From e66870e0569a8e803b4f968e39a567a54ab61444 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Tue, 14 Jul 2026 14:35:02 +1000 Subject: [PATCH 01/63] feat: limit the number of jobs submitted --- app/scheduler/jobs.py | 46 +++++++++++++++++++++- app/services/seqera.py | 22 ++++++++++- app/services/seqera_client.py | 3 ++ tests/scheduler/test_job_queue.py | 65 +++++++++++++++++++++++++++++++ 4 files changed, 133 insertions(+), 3 deletions(-) diff --git a/app/scheduler/jobs.py b/app/scheduler/jobs.py index 13895d8..939dc2d 100644 --- a/app/scheduler/jobs.py +++ b/app/scheduler/jobs.py @@ -1,4 +1,5 @@ import asyncio +import os from collections.abc import Awaitable from datetime import UTC, datetime, timedelta from typing import Protocol, cast @@ -11,16 +12,28 @@ from ..db.models.job_queue import QueuedJob from ..routes.dependencies import get_db from ..schemas.workflows import WorkflowName -from ..services import health +from ..services import health, seqera from ..services.bindflow_executor import launch_bindflow_workflow from ..services.proteinfold_executor import launch_proteinfold_workflow from ..services.seqera import WorkflowLaunchResult +from ..services.seqera_errors import SeqeraAPIError, SeqeraConfigurationError from ..services.wisps_executor import launch_wisps_workflow from . import SCHEDULER LAUNCH_MAX_ATTEMPTS = 3 RETRY_DELAY_BASE = 5 * 60 +# Gadi's gpuhopper PBS queue caps total concurrent jobs at GADI_JOB_LIMIT, and each +# submitted workflow holds up to GADI_QUEUE_SIZE_PER_WORKFLOW of those jobs for its +# whole run (Nextflow's queueSize), so at most job_limit / queue_size workflows can +# run at once (default: 50 / 2 = 25). MAX_CONCURRENT_WORKFLOWS can also be set +# directly to override that derived value. +GADI_JOB_LIMIT = int(os.getenv("GADI_JOB_LIMIT", "50")) +GADI_QUEUE_SIZE_PER_WORKFLOW = int(os.getenv("GADI_QUEUE_SIZE_PER_WORKFLOW", "2")) +MAX_CONCURRENT_WORKFLOWS = int( + os.getenv("MAX_CONCURRENT_WORKFLOWS", str(GADI_JOB_LIMIT // GADI_QUEUE_SIZE_PER_WORKFLOW)) +) + class LaunchFunction(Protocol): """ @@ -105,6 +118,20 @@ def launch_job(job_id: UUID, dry_run: bool = False) -> None: return +def get_available_workflow_capacity() -> int: + """ + How many more workflows can be submitted to Gadi right now, per the Seqera API's + count of workflows still occupying a job slot there (see MAX_CONCURRENT_WORKFLOWS). + """ + active_workflow_count = asyncio.run(seqera.count_active_workflows()) + capacity = max(0, MAX_CONCURRENT_WORKFLOWS - active_workflow_count) + logger.info( + f"{active_workflow_count}/{MAX_CONCURRENT_WORKFLOWS} workflows active on Gadi " + f"({capacity} submission slot(s) available)." + ) + return capacity + + def submit_pending_jobs(dry_run: bool = False): # Time between jobs job_offset = 10 @@ -115,6 +142,15 @@ def submit_pending_jobs(dry_run: bool = False): logger.warning("Skipping pending job submission while system status is unhealthy.") return + try: + available_capacity = get_available_workflow_capacity() + except (SeqeraAPIError, SeqeraConfigurationError) as e: + logger.warning(f"Could not determine Gadi workflow capacity from Seqera: {e}") + return + if available_capacity <= 0: + logger.info("Gadi is at its concurrent workflow limit; skipping submission this tick.") + return + now = datetime.now(tz=UTC) pending_query = select(QueuedJob).where( @@ -123,7 +159,13 @@ def submit_pending_jobs(dry_run: bool = False): pending_jobs = db_session.scalars(pending_query).all() logger.info(f"Found {len(pending_jobs)} pending jobs.") - for index, job in enumerate(pending_jobs): + jobs_to_submit = pending_jobs[:available_capacity] + if len(jobs_to_submit) < len(pending_jobs): + logger.info( + f"Only submitting {len(jobs_to_submit)} of {len(pending_jobs)} pending jobs " + "due to available Gadi capacity." + ) + for index, job in enumerate(jobs_to_submit): launch_id = f"launch_job_{job.id}" # Ignore if already scheduled if SCHEDULER.get_job(launch_id, jobstore="memory") is not None: diff --git a/app/services/seqera.py b/app/services/seqera.py index 53c3f54..4c2cef5 100644 --- a/app/services/seqera.py +++ b/app/services/seqera.py @@ -10,11 +10,20 @@ import httpx import yaml -from .seqera_client import SeqeraClient +from .seqera_client import SeqeraClient, list_workflows_raw from .seqera_errors import SeqeraAPIError, SeqeraConfigurationError +from .seqera_parsers import parse_workflow_list_payload logger = logging.getLogger(__name__) +# Pipeline statuses that still occupy a slot on Gadi (queued or actively running). +ACTIVE_PIPELINE_STATUSES = frozenset({"SUBMITTED", "RUNNING"}) + +# 100 is the Seqera API's hard ceiling for this endpoint's `max` param (a higher +# value gets a 400). The list is returned most-recent-first, so as long as active +# runs stay under 100 they're guaranteed to be on this single page - no paging. +_ACTIVE_WORKFLOW_LIST_MAX = 100 + class WorkflowExecutorError(RuntimeError): """Raised when workflow execution through Seqera fails.""" @@ -73,6 +82,17 @@ async def post_seqera_launch( ) +async def count_active_workflows(workspace_id: str | None = None) -> int: + """Count Seqera workflow runs that are still queued or running on Gadi. + + Used to cap how many new workflows the scheduler submits at once, since each + run holds a slice of Gadi's shared PBS job-slot limit for its whole lifetime. + """ + data = await list_workflows_raw(workspace_id, max_results=_ACTIVE_WORKFLOW_LIST_MAX) + items, _total = parse_workflow_list_payload(data) + return sum(1 for item in items if item.pipeline_status in ACTIVE_PIPELINE_STATUSES) + + def _get_required_env(key: str) -> str: """Get required environment variable or raise error.""" value = os.getenv(key) diff --git a/app/services/seqera_client.py b/app/services/seqera_client.py index 657a4ac..5c2f5dd 100644 --- a/app/services/seqera_client.py +++ b/app/services/seqera_client.py @@ -85,10 +85,13 @@ def _headers(token: str) -> dict[str, str]: async def list_workflows_raw( workspace_id: str | None = None, search_query: str | None = None, + max_results: int | None = None, ) -> dict[str, Any] | list[Any]: api_url, token, params = _get_api_context(workspace_id) if search_query: params["search"] = search_query + if max_results is not None: + params["max"] = max_results url = f"{api_url}/workflow" async with httpx.AsyncClient(timeout=httpx.Timeout(60)) as client: diff --git a/tests/scheduler/test_job_queue.py b/tests/scheduler/test_job_queue.py index 67ba32b..b1517d9 100644 --- a/tests/scheduler/test_job_queue.py +++ b/tests/scheduler/test_job_queue.py @@ -73,6 +73,9 @@ def test_submit_pending_jobs_schedules_only_due_pending_jobs( monkeypatch.setattr(scheduler_jobs, "get_db", _get_db_override(test_db)) monkeypatch.setattr(scheduler_jobs, "SCHEDULER", scheduler) monkeypatch.setattr(scheduler_jobs, "is_seqera_available", lambda _db_session: True) + monkeypatch.setattr( + scheduler_jobs, "get_available_workflow_capacity", lambda: scheduler_jobs.MAX_CONCURRENT_WORKFLOWS + ) scheduler_jobs.submit_pending_jobs(dry_run=True) @@ -103,9 +106,71 @@ def test_submit_pending_jobs_skips_jobs_already_scheduled(test_db, persistent_mo monkeypatch.setattr(scheduler_jobs, "get_db", _get_db_override(test_db)) monkeypatch.setattr(scheduler_jobs, "SCHEDULER", scheduler) monkeypatch.setattr(scheduler_jobs, "is_seqera_available", lambda _db_session: True) + monkeypatch.setattr( + scheduler_jobs, "get_available_workflow_capacity", lambda: scheduler_jobs.MAX_CONCURRENT_WORKFLOWS + ) scheduler_jobs.submit_pending_jobs() scheduled_jobs = scheduler.get_jobs(jobstore="memory") assert [job.id for job in scheduled_jobs] == [launch_id] assert scheduled_jobs[0].kwargs == {"job_id": due_job.id, "dry_run": False} + + +def test_submit_pending_jobs_skips_when_no_gadi_capacity( + test_db, persistent_models, monkeypatch +): + _create_queued_job(next_attempt_at=datetime.now(UTC) - timedelta(minutes=1)) + scheduler = _make_scheduler() + + monkeypatch.setattr(scheduler_jobs, "get_db", _get_db_override(test_db)) + monkeypatch.setattr(scheduler_jobs, "SCHEDULER", scheduler) + monkeypatch.setattr(scheduler_jobs, "is_seqera_available", lambda _db_session: True) + monkeypatch.setattr(scheduler_jobs, "get_available_workflow_capacity", lambda: 0) + + scheduler_jobs.submit_pending_jobs() + + assert scheduler.get_jobs(jobstore="memory") == [] + + +def test_submit_pending_jobs_caps_submissions_to_available_capacity( + test_db, persistent_models, monkeypatch +): + now = datetime.now(UTC) + due_jobs = [ + _create_queued_job(status="pending", next_attempt_at=now - timedelta(minutes=1)) + for _ in range(3) + ] + scheduler = _make_scheduler() + + monkeypatch.setattr(scheduler_jobs, "get_db", _get_db_override(test_db)) + monkeypatch.setattr(scheduler_jobs, "SCHEDULER", scheduler) + monkeypatch.setattr(scheduler_jobs, "is_seqera_available", lambda _db_session: True) + monkeypatch.setattr(scheduler_jobs, "get_available_workflow_capacity", lambda: 2) + + scheduler_jobs.submit_pending_jobs() + + scheduled_jobs = scheduler.get_jobs(jobstore="memory") + assert len(scheduled_jobs) == 2 + scheduled_job_ids = {job.kwargs["job_id"] for job in scheduled_jobs} + assert scheduled_job_ids.issubset({job.id for job in due_jobs}) + + +def test_get_available_workflow_capacity_uses_seqera_active_count(monkeypatch): + async def _count_active_workflows(): + return 20 + + monkeypatch.setattr(scheduler_jobs.seqera, "count_active_workflows", _count_active_workflows) + monkeypatch.setattr(scheduler_jobs, "MAX_CONCURRENT_WORKFLOWS", 25) + + assert scheduler_jobs.get_available_workflow_capacity() == 5 + + +def test_get_available_workflow_capacity_floors_at_zero(monkeypatch): + async def _count_active_workflows(): + return 30 + + monkeypatch.setattr(scheduler_jobs.seqera, "count_active_workflows", _count_active_workflows) + monkeypatch.setattr(scheduler_jobs, "MAX_CONCURRENT_WORKFLOWS", 25) + + assert scheduler_jobs.get_available_workflow_capacity() == 0 From c9071491ac7b814888f742c596827700bba6c595 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Tue, 14 Jul 2026 14:46:35 +1000 Subject: [PATCH 02/63] ci: update main to dev in CI workflows --- .github/workflows/check-migrations.yml | 2 +- .github/workflows/lint.yml | 4 ++-- .github/workflows/test-coverage.yml | 4 ++-- 3 files changed, 5 insertions(+), 5 deletions(-) diff --git a/.github/workflows/check-migrations.yml b/.github/workflows/check-migrations.yml index bb1ca37..29964d9 100644 --- a/.github/workflows/check-migrations.yml +++ b/.github/workflows/check-migrations.yml @@ -2,7 +2,7 @@ name: Check Alembic Migrations on: pull_request: - branches: [main] + branches: [dev] paths: - "alembic/**" - "alembic.ini" diff --git a/.github/workflows/lint.yml b/.github/workflows/lint.yml index ac039d5..4cb292b 100644 --- a/.github/workflows/lint.yml +++ b/.github/workflows/lint.yml @@ -2,9 +2,9 @@ name: Lint on: push: - branches: [main] + branches: [dev] pull_request: - branches: [main] + branches: [dev] jobs: ruff: diff --git a/.github/workflows/test-coverage.yml b/.github/workflows/test-coverage.yml index 3632112..00dcc73 100644 --- a/.github/workflows/test-coverage.yml +++ b/.github/workflows/test-coverage.yml @@ -2,9 +2,9 @@ name: Coverage on: push: - branches: ["main"] + branches: ["dev"] pull_request: - branches: ["main"] + branches: ["dev"] jobs: tests: From 13c6dfee2a22092eac4b2565ae9fefa3f68c5e10 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Tue, 14 Jul 2026 15:09:44 +1000 Subject: [PATCH 03/63] test: seqera services and job queue tests --- app/services/seqera_client.py | 4 +- tests/scheduler/test_job_queue.py | 15 ++++--- tests/test_services_seqera.py | 60 +++++++++++++++++++++++++++- tests/test_services_seqera_client.py | 41 +++++++++++++++++++ 4 files changed, 110 insertions(+), 10 deletions(-) diff --git a/app/services/seqera_client.py b/app/services/seqera_client.py index 5c2f5dd..355e9b4 100644 --- a/app/services/seqera_client.py +++ b/app/services/seqera_client.py @@ -65,11 +65,11 @@ def _get_required_env(key: str) -> str: return value -def _get_api_context(workspace_id: str | None = None) -> tuple[str, str, dict[str, str]]: +def _get_api_context(workspace_id: str | None = None) -> tuple[str, str, dict[str, Any]]: api_url = _get_required_env("SEQERA_API_URL").rstrip("/") token = _get_required_env("SEQERA_ACCESS_TOKEN") resolved_workspace = workspace_id or os.getenv("WORK_SPACE") - params: dict[str, str] = {} + params: dict[str, Any] = {} if resolved_workspace: params["workspaceId"] = resolved_workspace return api_url, token, params diff --git a/tests/scheduler/test_job_queue.py b/tests/scheduler/test_job_queue.py index b1517d9..8b3f7ba 100644 --- a/tests/scheduler/test_job_queue.py +++ b/tests/scheduler/test_job_queue.py @@ -1,4 +1,5 @@ from datetime import UTC, datetime, timedelta +from uuid import uuid4 from apscheduler.jobstores.memory import MemoryJobStore from apscheduler.schedulers.blocking import BlockingScheduler @@ -24,7 +25,7 @@ def _create_queued_job(*, status: str = "pending", next_attempt_at: datetime | N workflow_run = WorkflowRunFactory.create_sync( owner=user, workflow=workflow, - work_dir=f"/work/{status}-{next_attempt_at.timestamp() if next_attempt_at else 'none'}", + work_dir=f"/work/{status}-{next_attempt_at.timestamp() if next_attempt_at else 'none'}-{uuid4()}", ) return QueuedJobFactory.create_sync( workflow=workflow, @@ -74,7 +75,9 @@ def test_submit_pending_jobs_schedules_only_due_pending_jobs( monkeypatch.setattr(scheduler_jobs, "SCHEDULER", scheduler) monkeypatch.setattr(scheduler_jobs, "is_seqera_available", lambda _db_session: True) monkeypatch.setattr( - scheduler_jobs, "get_available_workflow_capacity", lambda: scheduler_jobs.MAX_CONCURRENT_WORKFLOWS + scheduler_jobs, + "get_available_workflow_capacity", + lambda: scheduler_jobs.MAX_CONCURRENT_WORKFLOWS, ) scheduler_jobs.submit_pending_jobs(dry_run=True) @@ -107,7 +110,9 @@ def test_submit_pending_jobs_skips_jobs_already_scheduled(test_db, persistent_mo monkeypatch.setattr(scheduler_jobs, "SCHEDULER", scheduler) monkeypatch.setattr(scheduler_jobs, "is_seqera_available", lambda _db_session: True) monkeypatch.setattr( - scheduler_jobs, "get_available_workflow_capacity", lambda: scheduler_jobs.MAX_CONCURRENT_WORKFLOWS + scheduler_jobs, + "get_available_workflow_capacity", + lambda: scheduler_jobs.MAX_CONCURRENT_WORKFLOWS, ) scheduler_jobs.submit_pending_jobs() @@ -117,9 +122,7 @@ def test_submit_pending_jobs_skips_jobs_already_scheduled(test_db, persistent_mo assert scheduled_jobs[0].kwargs == {"job_id": due_job.id, "dry_run": False} -def test_submit_pending_jobs_skips_when_no_gadi_capacity( - test_db, persistent_models, monkeypatch -): +def test_submit_pending_jobs_skips_when_no_gadi_capacity(test_db, persistent_models, monkeypatch): _create_queued_job(next_attempt_at=datetime.now(UTC) - timedelta(minutes=1)) scheduler = _make_scheduler() diff --git a/tests/test_services_seqera.py b/tests/test_services_seqera.py index 6ad158c..e3ff890 100644 --- a/tests/test_services_seqera.py +++ b/tests/test_services_seqera.py @@ -19,8 +19,8 @@ launch_bindflow_workflow, prepare_bindflow_workflow, ) -from app.services.seqera import WorkflowExecutorError, WorkflowLaunchResult -from app.services.seqera_errors import SeqeraConfigurationError +from app.services.seqera import WorkflowExecutorError, WorkflowLaunchResult, count_active_workflows +from app.services.seqera_errors import SeqeraAPIError, SeqeraConfigurationError from tests.datagen import AppUserFactory, QueuedJobFactory, WorkflowFactory, WorkflowRunFactory _CONFIG_PATH = "/some/bindflow.config" @@ -343,3 +343,59 @@ async def test_launch_with_custom_params_text(persistent_models): assert "my_custom_param: 42" in params_text assert "another_param: test" in params_text + + +def _workflow_list_response(*statuses: str) -> dict: + return { + "workflows": [ + {"workflow": {"id": f"wf-{i}", "status": status}} for i, status in enumerate(statuses) + ], + "totalSize": len(statuses), + "hasMore": False, + } + + +@pytest.mark.asyncio +@respx.mock +async def test_count_active_workflows_counts_running_and_submitted(): + respx.get(url__regex=r".*/workflow(\?.*)?$").mock( + return_value=httpx.Response( + 200, + json=_workflow_list_response( + "RUNNING", "SUBMITTED", "SUCCEEDED", "FAILED", "CANCELLED", "UNKNOWN" + ), + ) + ) + + assert await count_active_workflows() == 2 + + +@pytest.mark.asyncio +@respx.mock +async def test_count_active_workflows_requests_the_api_max_page_size(): + route = respx.get(url__regex=r".*/workflow(\?.*)?$").mock( + return_value=httpx.Response(200, json=_workflow_list_response()) + ) + + await count_active_workflows() + + assert route.calls.last.request.url.params["max"] == "100" + + +@pytest.mark.asyncio +@respx.mock +async def test_count_active_workflows_no_active_runs(): + respx.get(url__regex=r".*/workflow(\?.*)?$").mock( + return_value=httpx.Response(200, json=_workflow_list_response("SUCCEEDED", "FAILED")) + ) + + assert await count_active_workflows() == 0 + + +@pytest.mark.asyncio +@respx.mock +async def test_count_active_workflows_raises_on_api_error(): + respx.get(url__regex=r".*/workflow(\?.*)?$").mock(return_value=httpx.Response(500, text="boom")) + + with pytest.raises(SeqeraAPIError): + await count_active_workflows() diff --git a/tests/test_services_seqera_client.py b/tests/test_services_seqera_client.py index 243374d..75be57b 100644 --- a/tests/test_services_seqera_client.py +++ b/tests/test_services_seqera_client.py @@ -91,6 +91,47 @@ async def test_describe_and_list_success(monkeypatch): assert await get_workflow_logs_raw("wf-1") == {"ok": True} +@pytest.mark.asyncio +async def test_list_workflows_raw_passes_max_and_search_params(monkeypatch): + monkeypatch.setenv("SEQERA_API_URL", "https://api.seqera.test") + monkeypatch.setenv("SEQERA_ACCESS_TOKEN", "token") + monkeypatch.setenv("WORK_SPACE", "ws-1") + + ok = AsyncMock(spec=httpx.Response) + ok.is_error = False + ok.json.return_value = {"workflows": []} + + with patch("httpx.AsyncClient.get", return_value=ok) as mock_get: + result = await list_workflows_raw(search_query="status:RUNNING", max_results=100) + + assert result == {"workflows": []} + mock_get.assert_awaited_once_with( + "https://api.seqera.test/workflow", + params={"workspaceId": "ws-1", "search": "status:RUNNING", "max": 100}, + headers={"Authorization": "Bearer token", "Accept": "application/json"}, + ) + + +@pytest.mark.asyncio +async def test_list_workflows_raw_omits_max_when_not_given(monkeypatch): + monkeypatch.setenv("SEQERA_API_URL", "https://api.seqera.test") + monkeypatch.setenv("SEQERA_ACCESS_TOKEN", "token") + monkeypatch.delenv("WORK_SPACE", raising=False) + + ok = AsyncMock(spec=httpx.Response) + ok.is_error = False + ok.json.return_value = {"workflows": []} + + with patch("httpx.AsyncClient.get", return_value=ok) as mock_get: + await list_workflows_raw() + + mock_get.assert_awaited_once_with( + "https://api.seqera.test/workflow", + params={}, + headers={"Authorization": "Bearer token", "Accept": "application/json"}, + ) + + @pytest.mark.asyncio async def test_cancel_and_delete_paths(monkeypatch): monkeypatch.setenv("SEQERA_API_URL", "https://api.seqera.test") From 555f9a64bb1d7a638ab17e5f177439afe0f4c144 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Tue, 14 Jul 2026 15:18:50 +1000 Subject: [PATCH 04/63] fix: hardcode max concurrent workflows from now --- app/scheduler/jobs.py | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) diff --git a/app/scheduler/jobs.py b/app/scheduler/jobs.py index 939dc2d..233abf7 100644 --- a/app/scheduler/jobs.py +++ b/app/scheduler/jobs.py @@ -23,16 +23,11 @@ LAUNCH_MAX_ATTEMPTS = 3 RETRY_DELAY_BASE = 5 * 60 -# Gadi's gpuhopper PBS queue caps total concurrent jobs at GADI_JOB_LIMIT, and each -# submitted workflow holds up to GADI_QUEUE_SIZE_PER_WORKFLOW of those jobs for its -# whole run (Nextflow's queueSize), so at most job_limit / queue_size workflows can -# run at once (default: 50 / 2 = 25). MAX_CONCURRENT_WORKFLOWS can also be set -# directly to override that derived value. -GADI_JOB_LIMIT = int(os.getenv("GADI_JOB_LIMIT", "50")) -GADI_QUEUE_SIZE_PER_WORKFLOW = int(os.getenv("GADI_QUEUE_SIZE_PER_WORKFLOW", "2")) -MAX_CONCURRENT_WORKFLOWS = int( - os.getenv("MAX_CONCURRENT_WORKFLOWS", str(GADI_JOB_LIMIT // GADI_QUEUE_SIZE_PER_WORKFLOW)) -) +# Gadi's gpuhopper PBS queue holds 50 job slots and each workflow run occupies approximately 2 of +# them (Nextflow's queueSize), so 25 workflows can run concurrently. Hardcoded as a +# temporary MVP value; post-MVP this should be derived from real-time queue +# capacity instead of a fixed constant. +MAX_CONCURRENT_WORKFLOWS = int(os.getenv("MAX_CONCURRENT_WORKFLOWS", "25")) class LaunchFunction(Protocol): From 14b382c082e1ffb550be35664d3a44d83bee80db Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Wed, 15 Jul 2026 11:10:46 +1000 Subject: [PATCH 05/63] docs: update MAX_CONCURRENT_WORKFLOWS docs string --- app/scheduler/jobs.py | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/app/scheduler/jobs.py b/app/scheduler/jobs.py index 233abf7..177821c 100644 --- a/app/scheduler/jobs.py +++ b/app/scheduler/jobs.py @@ -25,8 +25,7 @@ # Gadi's gpuhopper PBS queue holds 50 job slots and each workflow run occupies approximately 2 of # them (Nextflow's queueSize), so 25 workflows can run concurrently. Hardcoded as a -# temporary MVP value; post-MVP this should be derived from real-time queue -# capacity instead of a fixed constant. +# temporary MVP value, configurable via env var. MAX_CONCURRENT_WORKFLOWS = int(os.getenv("MAX_CONCURRENT_WORKFLOWS", "25")) From dca5839619c78a026713a5430cfa183fbbb555c0 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Wed, 15 Jul 2026 11:39:02 +1000 Subject: [PATCH 06/63] fix: count active workflows via totalSize --- app/services/seqera.py | 23 +++++++++------ app/services/seqera_client.py | 6 ++-- tests/test_services_seqera.py | 43 ++++++++++++++-------------- tests/test_services_seqera_client.py | 2 +- 4 files changed, 39 insertions(+), 35 deletions(-) diff --git a/app/services/seqera.py b/app/services/seqera.py index 4c2cef5..1eec212 100644 --- a/app/services/seqera.py +++ b/app/services/seqera.py @@ -12,18 +12,12 @@ from .seqera_client import SeqeraClient, list_workflows_raw from .seqera_errors import SeqeraAPIError, SeqeraConfigurationError -from .seqera_parsers import parse_workflow_list_payload logger = logging.getLogger(__name__) # Pipeline statuses that still occupy a slot on Gadi (queued or actively running). ACTIVE_PIPELINE_STATUSES = frozenset({"SUBMITTED", "RUNNING"}) -# 100 is the Seqera API's hard ceiling for this endpoint's `max` param (a higher -# value gets a 400). The list is returned most-recent-first, so as long as active -# runs stay under 100 they're guaranteed to be on this single page - no paging. -_ACTIVE_WORKFLOW_LIST_MAX = 100 - class WorkflowExecutorError(RuntimeError): """Raised when workflow execution through Seqera fails.""" @@ -87,10 +81,21 @@ async def count_active_workflows(workspace_id: str | None = None) -> int: Used to cap how many new workflows the scheduler submits at once, since each run holds a slice of Gadi's shared PBS job-slot limit for its whole lifetime. + + Queries each active status separately via the API's `search=status:` + filter and reads the server-computed `totalSize`, rather than fetching a page + of workflows and counting client-side - `totalSize` reflects the full filtered + count regardless of the page size, so this is exact even if there are more + matching runs than fit on one page. """ - data = await list_workflows_raw(workspace_id, max_results=_ACTIVE_WORKFLOW_LIST_MAX) - items, _total = parse_workflow_list_payload(data) - return sum(1 for item in items if item.pipeline_status in ACTIVE_PIPELINE_STATUSES) + total = 0 + for status in ACTIVE_PIPELINE_STATUSES: + data = await list_workflows_raw( + workspace_id, search_query=f"status:{status}", max_results=1 + ) + if isinstance(data, dict): + total += data.get("totalSize") or 0 + return total def _get_required_env(key: str) -> str: diff --git a/app/services/seqera_client.py b/app/services/seqera_client.py index 355e9b4..9d9816b 100644 --- a/app/services/seqera_client.py +++ b/app/services/seqera_client.py @@ -65,11 +65,11 @@ def _get_required_env(key: str) -> str: return value -def _get_api_context(workspace_id: str | None = None) -> tuple[str, str, dict[str, Any]]: +def _get_api_context(workspace_id: str | None = None) -> tuple[str, str, dict[str, str]]: api_url = _get_required_env("SEQERA_API_URL").rstrip("/") token = _get_required_env("SEQERA_ACCESS_TOKEN") resolved_workspace = workspace_id or os.getenv("WORK_SPACE") - params: dict[str, Any] = {} + params: dict[str, str] = {} if resolved_workspace: params["workspaceId"] = resolved_workspace return api_url, token, params @@ -91,7 +91,7 @@ async def list_workflows_raw( if search_query: params["search"] = search_query if max_results is not None: - params["max"] = max_results + params["max"] = str(max_results) url = f"{api_url}/workflow" async with httpx.AsyncClient(timeout=httpx.Timeout(60)) as client: diff --git a/tests/test_services_seqera.py b/tests/test_services_seqera.py index e3ff890..f67c7a2 100644 --- a/tests/test_services_seqera.py +++ b/tests/test_services_seqera.py @@ -345,49 +345,48 @@ async def test_launch_with_custom_params_text(persistent_models): assert "another_param: test" in params_text -def _workflow_list_response(*statuses: str) -> dict: - return { - "workflows": [ - {"workflow": {"id": f"wf-{i}", "status": status}} for i, status in enumerate(statuses) - ], - "totalSize": len(statuses), - "hasMore": False, - } +def _totals_by_status_handler(totals: dict[str, int]): + """Fake /workflow endpoint: returns totalSize for whichever status: was searched.""" + + def _handler(request: httpx.Request) -> httpx.Response: + search = request.url.params.get("search", "") + status = search.removeprefix("status:") + return httpx.Response( + 200, json={"workflows": [], "totalSize": totals.get(status, 0), "hasMore": False} + ) + + return _handler @pytest.mark.asyncio @respx.mock -async def test_count_active_workflows_counts_running_and_submitted(): +async def test_count_active_workflows_sums_running_and_submitted_totals(): respx.get(url__regex=r".*/workflow(\?.*)?$").mock( - return_value=httpx.Response( - 200, - json=_workflow_list_response( - "RUNNING", "SUBMITTED", "SUCCEEDED", "FAILED", "CANCELLED", "UNKNOWN" - ), - ) + side_effect=_totals_by_status_handler({"RUNNING": 3, "SUBMITTED": 2}) ) - assert await count_active_workflows() == 2 + assert await count_active_workflows() == 5 @pytest.mark.asyncio @respx.mock -async def test_count_active_workflows_requests_the_api_max_page_size(): +async def test_count_active_workflows_queries_each_status_with_a_minimal_page(): route = respx.get(url__regex=r".*/workflow(\?.*)?$").mock( - return_value=httpx.Response(200, json=_workflow_list_response()) + side_effect=_totals_by_status_handler({}) ) await count_active_workflows() - assert route.calls.last.request.url.params["max"] == "100" + assert route.call_count == 2 + searches = {call.request.url.params["search"] for call in route.calls} + assert searches == {"status:RUNNING", "status:SUBMITTED"} + assert all(call.request.url.params["max"] == "1" for call in route.calls) @pytest.mark.asyncio @respx.mock async def test_count_active_workflows_no_active_runs(): - respx.get(url__regex=r".*/workflow(\?.*)?$").mock( - return_value=httpx.Response(200, json=_workflow_list_response("SUCCEEDED", "FAILED")) - ) + respx.get(url__regex=r".*/workflow(\?.*)?$").mock(side_effect=_totals_by_status_handler({})) assert await count_active_workflows() == 0 diff --git a/tests/test_services_seqera_client.py b/tests/test_services_seqera_client.py index 75be57b..9ffd1fc 100644 --- a/tests/test_services_seqera_client.py +++ b/tests/test_services_seqera_client.py @@ -107,7 +107,7 @@ async def test_list_workflows_raw_passes_max_and_search_params(monkeypatch): assert result == {"workflows": []} mock_get.assert_awaited_once_with( "https://api.seqera.test/workflow", - params={"workspaceId": "ws-1", "search": "status:RUNNING", "max": 100}, + params={"workspaceId": "ws-1", "search": "status:RUNNING", "max": "100"}, headers={"Authorization": "Bearer token", "Accept": "application/json"}, ) From c1a7cc955e805dc2dd39c9263f1abdf50a882407 Mon Sep 17 00:00:00 2001 From: laclac102 <53896516+vtnphan@users.noreply.github.com> Date: Wed, 15 Jul 2026 16:46:36 +1000 Subject: [PATCH 07/63] feat: base64 encode utils --- app/services/cluster_utils.py | 10 ++++++++++ 1 file changed, 10 insertions(+) create mode 100644 app/services/cluster_utils.py diff --git a/app/services/cluster_utils.py b/app/services/cluster_utils.py new file mode 100644 index 0000000..9e95f88 --- /dev/null +++ b/app/services/cluster_utils.py @@ -0,0 +1,10 @@ +"""Utilities for building HPC cluster option strings.""" + +from __future__ import annotations + +import base64 + + +def encode_ip(ip_address: str) -> str: + """Return the base64 encoding of an IP address string.""" + return base64.b64encode(ip_address.encode()).decode() From b3d3717c6c077b996aec681a16a9c890b884800e Mon Sep 17 00:00:00 2001 From: laclac102 <53896516+vtnphan@users.noreply.github.com> Date: Wed, 15 Jul 2026 16:50:17 +1000 Subject: [PATCH 08/63] feat: update cluster options --- app/services/bindflow_config.py | 14 +++++--------- app/services/bindflow_executor.py | 5 ++--- app/services/proteinfold_config.py | 14 +++++--------- app/services/proteinfold_executor.py | 5 ++--- app/services/wisps_config.py | 14 +++++--------- app/services/wisps_executor.py | 6 ++---- 6 files changed, 21 insertions(+), 37 deletions(-) diff --git a/app/services/bindflow_config.py b/app/services/bindflow_config.py index 750bde2..1468f62 100644 --- a/app/services/bindflow_config.py +++ b/app/services/bindflow_config.py @@ -4,7 +4,7 @@ from typing import Any -from ..schemas.workflows import WorkflowUserDetails +from .cluster_utils import encode_ip from .workflow_config_fetcher import fetch_workflow_config @@ -25,17 +25,13 @@ def get_bindflow_config_profiles() -> list[str]: def get_bindflow_config_text( config_file_path: str, *, - job_id: str, - user_details: WorkflowUserDetails, - timestamp: str, + email: str, + ip_address: str = "", ) -> str: """Read bindflow base config and append a process override block with runtime values.""" base = fetch_workflow_config(config_file_path) - cluster_opts = ( - f"-v JOB_ID={job_id},USER_NAME={user_details.user_email}," - f"TIMESTAMP={timestamp},FULL_NAME={user_details.full_name}," - f"INSTITUTE={user_details.institute},IP_ADDRESS={user_details.ip_address}" - ) + account = f"{email}:{encode_ip(ip_address)}" if ip_address else email + cluster_opts = f"-A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override diff --git a/app/services/bindflow_executor.py b/app/services/bindflow_executor.py index 9296160..0c88155 100644 --- a/app/services/bindflow_executor.py +++ b/app/services/bindflow_executor.py @@ -91,9 +91,8 @@ async def prepare_bindflow_workflow( # pylint: disable=too-many-locals "configProfiles": get_bindflow_config_profiles(), "configText": get_bindflow_config_text( config_path, - job_id=run_name, - user_details=user_details, - timestamp=timestamp, + email=user_details.user_email, + ip_address=user_details.ip_address, ), "resume": False, } diff --git a/app/services/proteinfold_config.py b/app/services/proteinfold_config.py index 8843a13..2410395 100644 --- a/app/services/proteinfold_config.py +++ b/app/services/proteinfold_config.py @@ -4,7 +4,7 @@ from typing import Any -from ..schemas.workflows import WorkflowUserDetails +from .cluster_utils import encode_ip from .workflow_config_fetcher import fetch_workflow_config @@ -23,17 +23,13 @@ def get_proteinfold_config_profiles() -> list[str]: def get_proteinfold_config_text( config_file_path: str, *, - job_id: str, - user_details: WorkflowUserDetails, - timestamp: str, + email: str, + ip_address: str = "", ) -> str: """Read proteinfold base config and append a process override block with runtime values.""" base = fetch_workflow_config(config_file_path) - cluster_opts = ( - f"-P yz52 -v JOB_ID={job_id},USER_NAME={user_details.user_email}," - f"TIMESTAMP={timestamp},FULL_NAME={user_details.full_name}," - f"INSTITUTE={user_details.institute},IP_ADDRESS={user_details.ip_address}" - ) + account = f"{email}:{encode_ip(ip_address)}" if ip_address else email + cluster_opts = f"-P yz52 -A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override diff --git a/app/services/proteinfold_executor.py b/app/services/proteinfold_executor.py index cfb99cc..d1cb2c8 100644 --- a/app/services/proteinfold_executor.py +++ b/app/services/proteinfold_executor.py @@ -119,9 +119,8 @@ async def prepare_proteinfold_workflow( "configProfiles": get_proteinfold_config_profiles(), "configText": get_proteinfold_config_text( config_path, - job_id=job_id, - user_details=user_details, - timestamp=timestamp, + email=user_details.user_email, + ip_address=user_details.ip_address, ), "resume": False, } diff --git a/app/services/wisps_config.py b/app/services/wisps_config.py index e246c28..12dba13 100644 --- a/app/services/wisps_config.py +++ b/app/services/wisps_config.py @@ -4,7 +4,7 @@ from typing import Any, Literal -from ..schemas.workflows import WorkflowUserDetails +from .cluster_utils import encode_ip from .workflow_config_fetcher import fetch_workflow_config WispsMode = Literal["g1-g2", "manual"] @@ -39,9 +39,8 @@ def get_wisps_config_profiles() -> list[str]: def get_wisps_config_text( config_file_path: str, *, - job_id: str, - user_details: WorkflowUserDetails, - timestamp: str, + email: str, + ip_address: str = "", ) -> str: """Read wisps config and append a process override block with runtime values. @@ -50,10 +49,7 @@ def get_wisps_config_text( """ base = fetch_workflow_config(config_file_path) - cluster_opts = ( - f"-P yz52 -v JOB_ID={job_id},USER_NAME={user_details.user_email}," - f"TIMESTAMP={timestamp},FULL_NAME={user_details.full_name}," - f"INSTITUTE={user_details.institute},IP_ADDRESS={user_details.ip_address}" - ) + account = f"{email}:{encode_ip(ip_address)}" if ip_address else email + cluster_opts = f"-P yz52 -A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override diff --git a/app/services/wisps_executor.py b/app/services/wisps_executor.py index a0a4208..3b75891 100644 --- a/app/services/wisps_executor.py +++ b/app/services/wisps_executor.py @@ -54,7 +54,6 @@ async def prepare_wisps_workflow( raise SeqeraConfigurationError("Missing output identifier for workflow launch") out_dir = f"s3://{s3_bucket}/{output_id.strip()}" - timestamp = datetime.now(UTC).strftime("%Y%m%d_%H%M%S") job_id = (form.runName or "").strip() if not job_id: raise SeqeraConfigurationError("Missing run name for workflow launch") @@ -72,9 +71,8 @@ async def prepare_wisps_workflow( config_text = get_wisps_config_text( config_path, - job_id=job_id, - user_details=user_details, - timestamp=timestamp, + email=user_details.user_email, + ip_address=user_details.ip_address, ) launch_payload: dict[str, Any] = { From c6e4bca164be77d63403cd088b1d55c76744ddd4 Mon Sep 17 00:00:00 2001 From: laclac102 <53896516+vtnphan@users.noreply.github.com> Date: Wed, 15 Jul 2026 16:53:15 +1000 Subject: [PATCH 09/63] refactor: update project to GADI_PROJECT param --- app/services/bindflow_config.py | 4 ++-- app/services/cluster_utils.py | 3 +++ app/services/proteinfold_config.py | 6 +++--- app/services/wisps_config.py | 4 ++-- 4 files changed, 10 insertions(+), 7 deletions(-) diff --git a/app/services/bindflow_config.py b/app/services/bindflow_config.py index 1468f62..3a2f872 100644 --- a/app/services/bindflow_config.py +++ b/app/services/bindflow_config.py @@ -4,14 +4,14 @@ from typing import Any -from .cluster_utils import encode_ip +from .cluster_utils import GADI_PROJECT, encode_ip from .workflow_config_fetcher import fetch_workflow_config def get_bindflow_default_params(out_dir: str, samplesheet_url: str) -> dict[str, Any]: """Get default parameters for bindflow workflow.""" return { - "project": "yz52", + "project": GADI_PROJECT, "outdir": out_dir, "input": samplesheet_url, } diff --git a/app/services/cluster_utils.py b/app/services/cluster_utils.py index 9e95f88..6d0af78 100644 --- a/app/services/cluster_utils.py +++ b/app/services/cluster_utils.py @@ -3,6 +3,9 @@ from __future__ import annotations import base64 +import os + +GADI_PROJECT: str = os.getenv("GADI_PROJECT", "yz52") def encode_ip(ip_address: str) -> str: diff --git a/app/services/proteinfold_config.py b/app/services/proteinfold_config.py index 2410395..7398ace 100644 --- a/app/services/proteinfold_config.py +++ b/app/services/proteinfold_config.py @@ -4,7 +4,7 @@ from typing import Any -from .cluster_utils import encode_ip +from .cluster_utils import GADI_PROJECT, encode_ip from .workflow_config_fetcher import fetch_workflow_config @@ -12,7 +12,7 @@ def get_proteinfold_default_params( out_dir: str, samplesheet_url: str, mode: str = "alphafold2" ) -> dict[str, Any]: """Get default parameters for proteinfold workflow.""" - return {"input": samplesheet_url, "outdir": out_dir, "project": "yz52", "mode": mode} + return {"input": samplesheet_url, "outdir": out_dir, "project": GADI_PROJECT, "mode": mode} def get_proteinfold_config_profiles() -> list[str]: @@ -30,6 +30,6 @@ def get_proteinfold_config_text( base = fetch_workflow_config(config_file_path) account = f"{email}:{encode_ip(ip_address)}" if ip_address else email - cluster_opts = f"-P yz52 -A {account}" + cluster_opts = f"-P {GADI_PROJECT} -A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override diff --git a/app/services/wisps_config.py b/app/services/wisps_config.py index 12dba13..070f9eb 100644 --- a/app/services/wisps_config.py +++ b/app/services/wisps_config.py @@ -4,7 +4,7 @@ from typing import Any, Literal -from .cluster_utils import encode_ip +from .cluster_utils import GADI_PROJECT, encode_ip from .workflow_config_fetcher import fetch_workflow_config WispsMode = Literal["g1-g2", "manual"] @@ -50,6 +50,6 @@ def get_wisps_config_text( base = fetch_workflow_config(config_file_path) account = f"{email}:{encode_ip(ip_address)}" if ip_address else email - cluster_opts = f"-P yz52 -A {account}" + cluster_opts = f"-P {GADI_PROJECT} -A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override From 9ad2654c0e8a1a5a5c75c04f0a63e732dc18748a Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Thu, 16 Jul 2026 09:08:36 +1000 Subject: [PATCH 10/63] tests: fix cluster options in tests --- tests/test_proteinfold_coverage.py | 30 +++++++------ tests/test_services_bindflow_config.py | 59 +++++++++----------------- tests/test_services_cluster_utils.py | 42 ++++++++++++++++++ tests/test_wisps_coverage.py | 29 ++++++++----- 4 files changed, 99 insertions(+), 61 deletions(-) create mode 100644 tests/test_services_cluster_utils.py diff --git a/tests/test_proteinfold_coverage.py b/tests/test_proteinfold_coverage.py index ecd9d80..f390e2f 100644 --- a/tests/test_proteinfold_coverage.py +++ b/tests/test_proteinfold_coverage.py @@ -544,33 +544,39 @@ def test_get_proteinfold_config_text_appends_process_block(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_proteinfold_config_text( "/fake/proteinfold.config", - job_id="my-job", - user_details=_USER_DETAILS, - timestamp="20240101_120000", + email=_USER_DETAILS.user_email, + ip_address=_USER_DETAILS.ip_address, ) assert "process {" in result assert "clusterOptions" in result -def test_get_proteinfold_config_text_contains_job_fields(): +def test_get_proteinfold_config_text_contains_email_and_encoded_ip(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_proteinfold_config_text( "/fake/proteinfold.config", - job_id="my-job", - user_details=_USER_DETAILS, - timestamp="20240101_120000", + email=_USER_DETAILS.user_email, + ip_address=_USER_DETAILS.ip_address, ) - assert "my-job" in result assert "user@ex.com" in result - assert "20240101_120000" in result + assert "MS4yLjMuNA==" in result + + +def test_get_proteinfold_config_text_without_ip_uses_email_only(): + with patch("builtins.open", mock_open(read_data="base_config")): + result = get_proteinfold_config_text( + "/fake/proteinfold.config", + email=_USER_DETAILS.user_email, + ) + assert "-A user@ex.com" in result + assert ":" not in result.split("clusterOptions = ")[1] def test_get_proteinfold_config_text_contains_base_config(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_proteinfold_config_text( "/fake/proteinfold.config", - job_id="my-job", - user_details=_USER_DETAILS, - timestamp="20240101_120000", + email=_USER_DETAILS.user_email, + ip_address=_USER_DETAILS.ip_address, ) assert "base_config" in result diff --git a/tests/test_services_bindflow_config.py b/tests/test_services_bindflow_config.py index 3ff9f77..7a5a150 100644 --- a/tests/test_services_bindflow_config.py +++ b/tests/test_services_bindflow_config.py @@ -9,7 +9,6 @@ import pytest import respx -from app.schemas.workflows import WorkflowUserDetails from app.services.bindflow_config import ( get_bindflow_config_profiles, get_bindflow_config_text, @@ -18,12 +17,6 @@ from app.services.launch_payloads import get_executor_script _SHEET_URL = "https://api.seqera.test/workspaces/ws1/datasets/ds1/v/1/n/samplesheet.csv" -_USER_DETAILS = WorkflowUserDetails( - user_email="user@example.com", - full_name="", - institute="", - ip_address="", -) # ============================================================================= # Tests for get_bindflow_default_params() @@ -125,9 +118,7 @@ def test_get_bindflow_config_text_includes_base_config(): with patch("builtins.open", mock_open(read_data="base_config_content")): result = get_bindflow_config_text( "/fake/bindflow.config", - job_id="run-1", - user_details=_USER_DETAILS, - timestamp="20260507_120000", + email="user@example.com", ) assert "base_config_content" in result @@ -136,43 +127,39 @@ def test_get_bindflow_config_text_appends_process_block(): with patch("builtins.open", mock_open(read_data="")): result = get_bindflow_config_text( "/fake/bindflow.config", - job_id="run-1", - user_details=_USER_DETAILS, - timestamp="20260507_120000", + email="user@example.com", ) assert "process {" in result assert "clusterOptions" in result -def test_get_bindflow_config_text_interpolates_job_fields(): +def test_get_bindflow_config_text_interpolates_email(): + with patch("builtins.open", mock_open(read_data="")): + result = get_bindflow_config_text( + "/fake/bindflow.config", + email="alice@example.com", + ) + assert "-A alice@example.com" in result + + +def test_get_bindflow_config_text_without_ip_address_omits_encoding(): with patch("builtins.open", mock_open(read_data="")): result = get_bindflow_config_text( "/fake/bindflow.config", - job_id="my-run", - user_details=_USER_DETAILS.model_copy(update={"user_email": "alice@example.com"}), - timestamp="20260507_090000", + email="user@example.com", ) - assert "my-run" in result - assert "alice@example.com" in result - assert "20260507_090000" in result + assert "-A user@example.com" in result + assert ":" not in result.split("clusterOptions = ")[1] -def test_get_bindflow_config_text_optional_fields(): +def test_get_bindflow_config_text_with_ip_address_appends_encoded_ip(): with patch("builtins.open", mock_open(read_data="")): result = get_bindflow_config_text( "/fake/bindflow.config", - job_id="run-1", - user_details=WorkflowUserDetails( - user_email="user@example.com", - full_name="Alice Smith", - institute="BioCommons", - ip_address="1.2.3.4", - ), - timestamp="ts", + email="user@example.com", + ip_address="1.2.3.4", ) - assert "Alice Smith" in result - assert "BioCommons" in result - assert "1.2.3.4" in result + assert "-A user@example.com:MS4yLjMuNA==" in result def test_get_bindflow_config_text_url_fetching(): @@ -182,9 +169,7 @@ def test_get_bindflow_config_text_url_fetching(): ) result = get_bindflow_config_text( "https://raw.githubusercontent.com/org/repo/main/bindflow.config", - job_id="run-url", - user_details=_USER_DETAILS, - timestamp="20260507_120000", + email="user@example.com", ) assert "remote_base_config" in result assert "clusterOptions" in result @@ -198,7 +183,5 @@ def test_get_bindflow_config_text_url_error_raises(): with pytest.raises(httpx.HTTPStatusError): get_bindflow_config_text( "https://raw.githubusercontent.com/org/repo/main/bindflow.config", - job_id="run-1", - user_details=_USER_DETAILS, - timestamp="ts", + email="user@example.com", ) diff --git a/tests/test_services_cluster_utils.py b/tests/test_services_cluster_utils.py new file mode 100644 index 0000000..fad9453 --- /dev/null +++ b/tests/test_services_cluster_utils.py @@ -0,0 +1,42 @@ +"""Tests for cluster_utils.py.""" + +# pylint: disable=missing-function-docstring +from __future__ import annotations + +import base64 +import importlib + +from app.services import cluster_utils +from app.services.cluster_utils import encode_ip + + +def test_encode_ip_returns_base64_of_ip_string(): + assert encode_ip("1.2.3.4") == base64.b64encode(b"1.2.3.4").decode() + + +def test_encode_ip_empty_string(): + assert encode_ip("") == "" + + +def test_encode_ip_ipv6_address(): + ip = "2001:db8::1" + assert encode_ip(ip) == base64.b64encode(ip.encode()).decode() + + +def test_gadi_project_defaults_to_yz52(monkeypatch): + monkeypatch.delenv("GADI_PROJECT", raising=False) + reloaded = importlib.reload(cluster_utils) + try: + assert reloaded.GADI_PROJECT == "yz52" + finally: + importlib.reload(cluster_utils) + + +def test_gadi_project_reads_env_var(monkeypatch): + monkeypatch.setenv("GADI_PROJECT", "ab12") + reloaded = importlib.reload(cluster_utils) + try: + assert reloaded.GADI_PROJECT == "ab12" + finally: + monkeypatch.delenv("GADI_PROJECT", raising=False) + importlib.reload(cluster_utils) diff --git a/tests/test_wisps_coverage.py b/tests/test_wisps_coverage.py index d269ced..5f9c4c6 100644 --- a/tests/test_wisps_coverage.py +++ b/tests/test_wisps_coverage.py @@ -258,33 +258,40 @@ def test_get_wisps_config_text_appends_process_block(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_wisps_config_text( config_file_path="/fake/path.config", - job_id="my-job", - user_details=_USER_DETAILS, - timestamp="20240101_120000", + email=_USER_DETAILS.user_email, + ip_address=_USER_DETAILS.ip_address, ) assert "process {" in result assert "clusterOptions" in result -def test_get_wisps_config_text_contains_job_fields(): +def test_get_wisps_config_text_contains_email_and_encoded_ip(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_wisps_config_text( config_file_path="/fake/path.config", - job_id="my-job", - user_details=_USER_DETAILS, - timestamp="20240101_120000", + email=_USER_DETAILS.user_email, + ip_address=_USER_DETAILS.ip_address, ) - assert "my-job" in result assert "user@ex.com" in result + assert "MS4yLjMuNA==" in result + + +def test_get_wisps_config_text_without_ip_uses_email_only(): + with patch("builtins.open", mock_open(read_data="base_config")): + result = get_wisps_config_text( + config_file_path="/fake/path.config", + email=_USER_DETAILS.user_email, + ) + assert "-A user@ex.com" in result + assert ":" not in result.split("clusterOptions = ")[1] def test_get_wisps_config_text_contains_base_config(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_wisps_config_text( config_file_path="/fake/path.config", - job_id="my-job", - user_details=_USER_DETAILS, - timestamp="20240101_120000", + email=_USER_DETAILS.user_email, + ip_address=_USER_DETAILS.ip_address, ) assert "base_config" in result From b25f78f1d1760227e14af61e24c8122145d832e2 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Thu, 16 Jul 2026 10:28:14 +1000 Subject: [PATCH 11/63] fix: add load dotenv to run_scheduler.py --- app/run_scheduler.py | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/app/run_scheduler.py b/app/run_scheduler.py index 5f1779e..35e9e1e 100644 --- a/app/run_scheduler.py +++ b/app/run_scheduler.py @@ -3,10 +3,13 @@ import typer from apscheduler.triggers.cron import CronTrigger from apscheduler.triggers.interval import IntervalTrigger +from dotenv import load_dotenv from loguru import logger -from app.scheduler import SCHEDULER -from app.scheduler.jobs import refresh_user_credits, submit_pending_jobs +load_dotenv() + +from app.scheduler import SCHEDULER # noqa: E402 +from app.scheduler.jobs import refresh_user_credits, submit_pending_jobs # noqa: E402 SUBMIT_INTERVAL = IntervalTrigger(minutes=5) MONTHLY_TRIGGER = CronTrigger(day=1, hour=1, minute=0, timezone="UTC") From ab7bd655b9d9251183f6b1e169bb502cd532c082 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Thu, 16 Jul 2026 11:37:41 +1000 Subject: [PATCH 12/63] fix: remove timestamp, username, jobid --- app/services/bindflow_executor.py | 5 ----- app/services/proteinfold_executor.py | 12 +----------- 2 files changed, 1 insertion(+), 16 deletions(-) diff --git a/app/services/bindflow_executor.py b/app/services/bindflow_executor.py index 0c88155..9697525 100644 --- a/app/services/bindflow_executor.py +++ b/app/services/bindflow_executor.py @@ -58,14 +58,9 @@ async def prepare_bindflow_workflow( # pylint: disable=too-many-locals raise SeqeraConfigurationError("Missing output identifier for workflow launch") out_dir = f"s3://{s3_bucket}/{output_key}" - timestamp = datetime.now(UTC).strftime("%Y%m%d_%H%M%S") - dataset_url = f"s3://{s3_bucket}/{s3_input_key}" default_params = get_bindflow_default_params(out_dir, dataset_url) - default_params["job_id"] = run_name - default_params["user_name"] = user_details.user_email - default_params["timestamp"] = timestamp default_params["mode"] = mode # Merge any tool-specific params forwarded from the frontend form diff --git a/app/services/proteinfold_executor.py b/app/services/proteinfold_executor.py index d1cb2c8..60ceaa5 100644 --- a/app/services/proteinfold_executor.py +++ b/app/services/proteinfold_executor.py @@ -50,14 +50,11 @@ def _build_params_text( mode: str, form_data: WorkflowFormData | None, custom_params: str | None, - extra_params: dict[str, Any] | None = None, ) -> str: """Build the YAML params string for the Seqera launch payload.""" params = get_proteinfold_default_params(out_dir, samplesheet_url, mode) if form_data: params.update(_tool_params(form_data)) - if extra_params: - params.update(extra_params) params_text = params_to_yaml_text(params) if custom_params and custom_params.strip(): params_text = f"{params_text}\n{custom_params.rstrip()}" @@ -89,9 +86,7 @@ async def prepare_proteinfold_workflow( raise SeqeraConfigurationError("Missing output identifier for workflow launch") out_dir = f"s3://{s3_bucket}/{output_id.strip()}" - timestamp = datetime.now(UTC).strftime("%Y%m%d_%H%M%S") - job_id = (form.runName or "").strip() - if not job_id: + if not form.runName or not form.runName.strip(): raise SeqeraConfigurationError("Missing run name for workflow launch") sheet_url = f"s3://{s3_bucket}/{s3_input_key}" @@ -101,11 +96,6 @@ async def prepare_proteinfold_workflow( mode, form_data, form.paramsText, - extra_params={ - "job_id": job_id, - "user_name": user_details.user_email, - "timestamp": timestamp, - }, ) launch_payload: dict[str, Any] = { From 3706861bc547bb725aea796cbb8e993d47586c14 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Thu, 16 Jul 2026 11:56:11 +1000 Subject: [PATCH 13/63] docs: add load env docs in app scheduler --- app/run_scheduler.py | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/app/run_scheduler.py b/app/run_scheduler.py index 35e9e1e..bd0d609 100644 --- a/app/run_scheduler.py +++ b/app/run_scheduler.py @@ -6,6 +6,10 @@ from dotenv import load_dotenv from loguru import logger +# Load .env before importing scheduler modules, which read required env vars (e.g. +# SEQERA_API_URL) at call time. Unlike app/main.py, this entrypoint runs standalone +# (`python -m app.run_scheduler`) and never imports main.py, so without this call +# .env values are never loaded into the process environment. load_dotenv() from app.scheduler import SCHEDULER # noqa: E402 From 12fa0c2ecd09b91e049fb1751fe54d903e01db8b Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Thu, 16 Jul 2026 11:57:32 +1000 Subject: [PATCH 14/63] docs: add GADI_PROJECT in .env.example --- .env.example | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/.env.example b/.env.example index 61cb94d..bfa9e7c 100644 --- a/.env.example +++ b/.env.example @@ -10,6 +10,10 @@ COMPUTE_ID=compute-env-id WORK_SPACE=workspace-id WORK_DIR=s3://bucket/workdir +# NCI Gadi project code used for PBS submissions (clusterOptions) across all +# Nextflow workflows. Optional; defaults to "yz52" if unset. +GADI_PROJECT=yz52 + # AWS S3 Configuration for file uploads AWS_ACCESS_KEY_ID=your-access-key-id AWS_SECRET_ACCESS_KEY=your-secret-access-key From 62a07bfa1b446ba760bcd626809b3c24155763de Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Thu, 16 Jul 2026 12:03:33 +1000 Subject: [PATCH 15/63] fix: remove unused user details --- app/routes/workflows.py | 21 ++++----------------- app/schemas/workflows.py | 2 -- pyproject.toml | 1 - uv.lock | 11 ----------- 4 files changed, 4 insertions(+), 31 deletions(-) diff --git a/app/routes/workflows.py b/app/routes/workflows.py index 5bf2e54..8500277 100644 --- a/app/routes/workflows.py +++ b/app/routes/workflows.py @@ -14,7 +14,6 @@ from pydantic import ValidationError from sqlalchemy import CursorResult, func, select, update from sqlalchemy.orm import Session -from unidecode import unidecode from ..db.models import QueuedJob from ..db.models.core import AppUser, RunInput, RunMetric, S3Object, Workflow, WorkflowRun @@ -234,28 +233,16 @@ async def launch_workflow( ) user = db_session.execute( - select(AppUser.email, AppUser.name).where(AppUser.id == current_user_id) + select(AppUser.email).where(AppUser.id == current_user_id) ).one_or_none() if not user or not user.email: raise HTTPException( status_code=status.HTTP_500_INTERNAL_SERVER_ERROR, detail="Could not retrieve user details required for workflow launch.", ) - user_email = user.email - # removes everything that isn't a letter, digit, or space - name = unidecode(user.name or "") - full_name = re.sub(r"[^a-zA-Z0-9 ]", "", name).replace(" ", "_") - institute = user_email.split("@")[-1] if "@" in user_email else None - ip_address: str | None = launch_ip or None - - full_name = _require_launch_var("full_name", full_name) - institute = _require_launch_var("institute", institute) - ip_address = _require_launch_var("ip_address", ip_address) user_details = WorkflowUserDetails( - user_email=user_email, - full_name=full_name, - institute=institute, - ip_address=ip_address, + user_email=user.email, + ip_address=_require_launch_var("ip_address", launch_ip or None), ) # Authoritative credit cost (server-side, non-spoofable). Only charged for @@ -406,7 +393,7 @@ async def launch_workflow( .values( credit=AppUser.credit - run_credit_cost, credit_updated_at=datetime.now(UTC), - credit_updated_by=user_email, + credit_updated_by=user_details.user_email, ) ), ) diff --git a/app/schemas/workflows.py b/app/schemas/workflows.py index 0dca971..a19370d 100644 --- a/app/schemas/workflows.py +++ b/app/schemas/workflows.py @@ -97,8 +97,6 @@ class WorkflowUserDetails(BaseModel): """ user_email: str = Field(..., description="Email address of the user") - full_name: str = Field(..., description="Full name of the user") - institute: str = Field(..., description="Institute of the user") ip_address: str = Field(..., description="IP address of the user") diff --git a/pyproject.toml b/pyproject.toml index 9ed45c7..91b0591 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -22,7 +22,6 @@ dependencies = [ "cachetools~=5.5", "python-jose[cryptography]~=3.5", "starlette-admin~=0.16", - "unidecode~=1.3", "apscheduler>=3.11.3", "loguru>=0.7.3", "typer>=0.26.8", diff --git a/uv.lock b/uv.lock index 14d44fb..ec6b5d1 100644 --- a/uv.lock +++ b/uv.lock @@ -1083,7 +1083,6 @@ dependencies = [ { name = "sqlalchemy" }, { name = "starlette-admin" }, { name = "typer" }, - { name = "unidecode" }, { name = "uvicorn", extra = ["standard"] }, ] @@ -1126,7 +1125,6 @@ requires-dist = [ { name = "sqlalchemy", specifier = "~=2.0" }, { name = "starlette-admin", specifier = "~=0.16" }, { name = "typer", specifier = ">=0.26.8" }, - { name = "unidecode", specifier = "~=1.3" }, { name = "uvicorn", extras = ["standard"], specifier = "~=0.49" }, ] @@ -1320,15 +1318,6 @@ wheels = [ { url = "https://files.pythonhosted.org/packages/9e/a4/017a7a6cbe387d961a688ec31364ae60a5c4e22c96ae9921b79a947c855d/tzlocal-5.4.4-py3-none-any.whl", hash = "sha256:aae09f0126a8a86fa736be266eb4a471380d26a0de3bc14844e7821fee3e2a15", size = 18115, upload-time = "2026-06-29T08:03:38.666Z" }, ] -[[package]] -name = "unidecode" -version = "1.4.0" -source = { registry = "https://pypi.org/simple" } -sdist = { url = "https://files.pythonhosted.org/packages/94/7d/a8a765761bbc0c836e397a2e48d498305a865b70a8600fd7a942e85dcf63/Unidecode-1.4.0.tar.gz", hash = "sha256:ce35985008338b676573023acc382d62c264f307c8f7963733405add37ea2b23", size = 200149, upload-time = "2025-04-24T08:45:03.798Z" } -wheels = [ - { url = "https://files.pythonhosted.org/packages/8f/b7/559f59d57d18b44c6d1250d2eeaa676e028b9c527431f5d0736478a73ba1/Unidecode-1.4.0-py3-none-any.whl", hash = "sha256:c3c7606c27503ad8d501270406e345ddb480a7b5f38827eafe4fa82a137f0021", size = 235837, upload-time = "2025-04-24T08:45:01.609Z" }, -] - [[package]] name = "urllib3" version = "2.7.0" From 90693ed79107e66d3e2b0bd7fa91b6dc9186e1fd Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Thu, 16 Jul 2026 12:03:50 +1000 Subject: [PATCH 16/63] fix: remove unused details in executor and config --- app/services/bindflow_config.py | 10 +++++++--- app/services/bindflow_executor.py | 3 +-- app/services/proteinfold_config.py | 10 +++++++--- app/services/proteinfold_executor.py | 3 +-- app/services/wisps_config.py | 10 +++++++--- app/services/wisps_executor.py | 3 +-- 6 files changed, 24 insertions(+), 15 deletions(-) diff --git a/app/services/bindflow_config.py b/app/services/bindflow_config.py index 3a2f872..359342f 100644 --- a/app/services/bindflow_config.py +++ b/app/services/bindflow_config.py @@ -4,6 +4,7 @@ from typing import Any +from ..schemas.workflows import WorkflowUserDetails from .cluster_utils import GADI_PROJECT, encode_ip from .workflow_config_fetcher import fetch_workflow_config @@ -25,13 +26,16 @@ def get_bindflow_config_profiles() -> list[str]: def get_bindflow_config_text( config_file_path: str, *, - email: str, - ip_address: str = "", + user_details: WorkflowUserDetails, ) -> str: """Read bindflow base config and append a process override block with runtime values.""" base = fetch_workflow_config(config_file_path) - account = f"{email}:{encode_ip(ip_address)}" if ip_address else email + account = ( + f"{user_details.user_email}:{encode_ip(user_details.ip_address)}" + if user_details.ip_address + else user_details.user_email + ) cluster_opts = f"-A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override diff --git a/app/services/bindflow_executor.py b/app/services/bindflow_executor.py index 9697525..a29ca05 100644 --- a/app/services/bindflow_executor.py +++ b/app/services/bindflow_executor.py @@ -86,8 +86,7 @@ async def prepare_bindflow_workflow( # pylint: disable=too-many-locals "configProfiles": get_bindflow_config_profiles(), "configText": get_bindflow_config_text( config_path, - email=user_details.user_email, - ip_address=user_details.ip_address, + user_details=user_details, ), "resume": False, } diff --git a/app/services/proteinfold_config.py b/app/services/proteinfold_config.py index 7398ace..4241c49 100644 --- a/app/services/proteinfold_config.py +++ b/app/services/proteinfold_config.py @@ -4,6 +4,7 @@ from typing import Any +from ..schemas.workflows import WorkflowUserDetails from .cluster_utils import GADI_PROJECT, encode_ip from .workflow_config_fetcher import fetch_workflow_config @@ -23,13 +24,16 @@ def get_proteinfold_config_profiles() -> list[str]: def get_proteinfold_config_text( config_file_path: str, *, - email: str, - ip_address: str = "", + user_details: WorkflowUserDetails, ) -> str: """Read proteinfold base config and append a process override block with runtime values.""" base = fetch_workflow_config(config_file_path) - account = f"{email}:{encode_ip(ip_address)}" if ip_address else email + account = ( + f"{user_details.user_email}:{encode_ip(user_details.ip_address)}" + if user_details.ip_address + else user_details.user_email + ) cluster_opts = f"-P {GADI_PROJECT} -A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override diff --git a/app/services/proteinfold_executor.py b/app/services/proteinfold_executor.py index 60ceaa5..59002fb 100644 --- a/app/services/proteinfold_executor.py +++ b/app/services/proteinfold_executor.py @@ -109,8 +109,7 @@ async def prepare_proteinfold_workflow( "configProfiles": get_proteinfold_config_profiles(), "configText": get_proteinfold_config_text( config_path, - email=user_details.user_email, - ip_address=user_details.ip_address, + user_details=user_details, ), "resume": False, } diff --git a/app/services/wisps_config.py b/app/services/wisps_config.py index 070f9eb..6f18c65 100644 --- a/app/services/wisps_config.py +++ b/app/services/wisps_config.py @@ -4,6 +4,7 @@ from typing import Any, Literal +from ..schemas.workflows import WorkflowUserDetails from .cluster_utils import GADI_PROJECT, encode_ip from .workflow_config_fetcher import fetch_workflow_config @@ -39,8 +40,7 @@ def get_wisps_config_profiles() -> list[str]: def get_wisps_config_text( config_file_path: str, *, - email: str, - ip_address: str = "", + user_details: WorkflowUserDetails, ) -> str: """Read wisps config and append a process override block with runtime values. @@ -49,7 +49,11 @@ def get_wisps_config_text( """ base = fetch_workflow_config(config_file_path) - account = f"{email}:{encode_ip(ip_address)}" if ip_address else email + account = ( + f"{user_details.user_email}:{encode_ip(user_details.ip_address)}" + if user_details.ip_address + else user_details.user_email + ) cluster_opts = f"-P {GADI_PROJECT} -A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override diff --git a/app/services/wisps_executor.py b/app/services/wisps_executor.py index 3b75891..7abe379 100644 --- a/app/services/wisps_executor.py +++ b/app/services/wisps_executor.py @@ -71,8 +71,7 @@ async def prepare_wisps_workflow( config_text = get_wisps_config_text( config_path, - email=user_details.user_email, - ip_address=user_details.ip_address, + user_details=user_details, ) launch_payload: dict[str, Any] = { From 5b24ac8ecd67d2b14685df7a4038523935c819f1 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Thu, 16 Jul 2026 12:04:49 +1000 Subject: [PATCH 17/63] test: fix related tests --- tests/test_proteinfold_coverage.py | 14 ++++---------- tests/test_services_bindflow_config.py | 22 ++++++++++++++-------- tests/test_services_seqera.py | 2 -- tests/test_wisps_coverage.py | 13 ++++--------- 4 files changed, 22 insertions(+), 29 deletions(-) diff --git a/tests/test_proteinfold_coverage.py b/tests/test_proteinfold_coverage.py index f390e2f..b96a77b 100644 --- a/tests/test_proteinfold_coverage.py +++ b/tests/test_proteinfold_coverage.py @@ -35,8 +35,6 @@ _USER_DETAILS = WorkflowUserDetails( user_email="user@ex.com", - full_name="Test_User", - institute="USYD", ip_address="1.2.3.4", ) @@ -369,7 +367,6 @@ async def test_prepare_proteinfold_workflow_writes_expected_queued_job( user_details=_USER_DETAILS.model_copy( update={ "user_email": "test@example.com", - "institute": "example.com", "ip_address": "127.0.0.1", } ), @@ -544,8 +541,7 @@ def test_get_proteinfold_config_text_appends_process_block(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_proteinfold_config_text( "/fake/proteinfold.config", - email=_USER_DETAILS.user_email, - ip_address=_USER_DETAILS.ip_address, + user_details=_USER_DETAILS, ) assert "process {" in result assert "clusterOptions" in result @@ -555,8 +551,7 @@ def test_get_proteinfold_config_text_contains_email_and_encoded_ip(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_proteinfold_config_text( "/fake/proteinfold.config", - email=_USER_DETAILS.user_email, - ip_address=_USER_DETAILS.ip_address, + user_details=_USER_DETAILS, ) assert "user@ex.com" in result assert "MS4yLjMuNA==" in result @@ -566,7 +561,7 @@ def test_get_proteinfold_config_text_without_ip_uses_email_only(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_proteinfold_config_text( "/fake/proteinfold.config", - email=_USER_DETAILS.user_email, + user_details=_USER_DETAILS.model_copy(update={"ip_address": ""}), ) assert "-A user@ex.com" in result assert ":" not in result.split("clusterOptions = ")[1] @@ -576,7 +571,6 @@ def test_get_proteinfold_config_text_contains_base_config(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_proteinfold_config_text( "/fake/proteinfold.config", - email=_USER_DETAILS.user_email, - ip_address=_USER_DETAILS.ip_address, + user_details=_USER_DETAILS, ) assert "base_config" in result diff --git a/tests/test_services_bindflow_config.py b/tests/test_services_bindflow_config.py index 7a5a150..09f1f46 100644 --- a/tests/test_services_bindflow_config.py +++ b/tests/test_services_bindflow_config.py @@ -9,6 +9,7 @@ import pytest import respx +from app.schemas.workflows import WorkflowUserDetails from app.services.bindflow_config import ( get_bindflow_config_profiles, get_bindflow_config_text, @@ -18,6 +19,12 @@ _SHEET_URL = "https://api.seqera.test/workspaces/ws1/datasets/ds1/v/1/n/samplesheet.csv" + +def _user_details(email: str, ip_address: str = "") -> WorkflowUserDetails: + return WorkflowUserDetails( + user_email=email, ip_address=ip_address + ) + # ============================================================================= # Tests for get_bindflow_default_params() # ============================================================================= @@ -118,7 +125,7 @@ def test_get_bindflow_config_text_includes_base_config(): with patch("builtins.open", mock_open(read_data="base_config_content")): result = get_bindflow_config_text( "/fake/bindflow.config", - email="user@example.com", + user_details=_user_details("user@example.com"), ) assert "base_config_content" in result @@ -127,7 +134,7 @@ def test_get_bindflow_config_text_appends_process_block(): with patch("builtins.open", mock_open(read_data="")): result = get_bindflow_config_text( "/fake/bindflow.config", - email="user@example.com", + user_details=_user_details("user@example.com"), ) assert "process {" in result assert "clusterOptions" in result @@ -137,7 +144,7 @@ def test_get_bindflow_config_text_interpolates_email(): with patch("builtins.open", mock_open(read_data="")): result = get_bindflow_config_text( "/fake/bindflow.config", - email="alice@example.com", + user_details=_user_details("alice@example.com"), ) assert "-A alice@example.com" in result @@ -146,7 +153,7 @@ def test_get_bindflow_config_text_without_ip_address_omits_encoding(): with patch("builtins.open", mock_open(read_data="")): result = get_bindflow_config_text( "/fake/bindflow.config", - email="user@example.com", + user_details=_user_details("user@example.com"), ) assert "-A user@example.com" in result assert ":" not in result.split("clusterOptions = ")[1] @@ -156,8 +163,7 @@ def test_get_bindflow_config_text_with_ip_address_appends_encoded_ip(): with patch("builtins.open", mock_open(read_data="")): result = get_bindflow_config_text( "/fake/bindflow.config", - email="user@example.com", - ip_address="1.2.3.4", + user_details=_user_details("user@example.com", ip_address="1.2.3.4"), ) assert "-A user@example.com:MS4yLjMuNA==" in result @@ -169,7 +175,7 @@ def test_get_bindflow_config_text_url_fetching(): ) result = get_bindflow_config_text( "https://raw.githubusercontent.com/org/repo/main/bindflow.config", - email="user@example.com", + user_details=_user_details("user@example.com"), ) assert "remote_base_config" in result assert "clusterOptions" in result @@ -183,5 +189,5 @@ def test_get_bindflow_config_text_url_error_raises(): with pytest.raises(httpx.HTTPStatusError): get_bindflow_config_text( "https://raw.githubusercontent.com/org/repo/main/bindflow.config", - email="user@example.com", + user_details=_user_details("user@example.com"), ) diff --git a/tests/test_services_seqera.py b/tests/test_services_seqera.py index f67c7a2..fad8757 100644 --- a/tests/test_services_seqera.py +++ b/tests/test_services_seqera.py @@ -26,8 +26,6 @@ _CONFIG_PATH = "/some/bindflow.config" _USER_DETAILS = WorkflowUserDetails( user_email="test@example.com", - full_name="Test_User", - institute="example.com", ip_address="127.0.0.1", ) diff --git a/tests/test_wisps_coverage.py b/tests/test_wisps_coverage.py index 5f9c4c6..f842c61 100644 --- a/tests/test_wisps_coverage.py +++ b/tests/test_wisps_coverage.py @@ -41,8 +41,6 @@ _USER_DETAILS = WorkflowUserDetails( user_email="user@ex.com", - full_name="Test User", - institute="USYD", ip_address="1.2.3.4", ) @@ -258,8 +256,7 @@ def test_get_wisps_config_text_appends_process_block(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_wisps_config_text( config_file_path="/fake/path.config", - email=_USER_DETAILS.user_email, - ip_address=_USER_DETAILS.ip_address, + user_details=_USER_DETAILS, ) assert "process {" in result assert "clusterOptions" in result @@ -269,8 +266,7 @@ def test_get_wisps_config_text_contains_email_and_encoded_ip(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_wisps_config_text( config_file_path="/fake/path.config", - email=_USER_DETAILS.user_email, - ip_address=_USER_DETAILS.ip_address, + user_details=_USER_DETAILS, ) assert "user@ex.com" in result assert "MS4yLjMuNA==" in result @@ -280,7 +276,7 @@ def test_get_wisps_config_text_without_ip_uses_email_only(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_wisps_config_text( config_file_path="/fake/path.config", - email=_USER_DETAILS.user_email, + user_details=_USER_DETAILS.model_copy(update={"ip_address": ""}), ) assert "-A user@ex.com" in result assert ":" not in result.split("clusterOptions = ")[1] @@ -290,8 +286,7 @@ def test_get_wisps_config_text_contains_base_config(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_wisps_config_text( config_file_path="/fake/path.config", - email=_USER_DETAILS.user_email, - ip_address=_USER_DETAILS.ip_address, + user_details=_USER_DETAILS, ) assert "base_config" in result From ffe7706ac67a2dcb0d9126fd3d655b4d69752bfe Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Thu, 16 Jul 2026 12:07:38 +1000 Subject: [PATCH 18/63] chore: lint --- tests/test_services_bindflow_config.py | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/tests/test_services_bindflow_config.py b/tests/test_services_bindflow_config.py index 09f1f46..95935c0 100644 --- a/tests/test_services_bindflow_config.py +++ b/tests/test_services_bindflow_config.py @@ -21,9 +21,8 @@ def _user_details(email: str, ip_address: str = "") -> WorkflowUserDetails: - return WorkflowUserDetails( - user_email=email, ip_address=ip_address - ) + return WorkflowUserDetails(user_email=email, ip_address=ip_address) + # ============================================================================= # Tests for get_bindflow_default_params() From ffd4b892f3235366d230b168c69f93c7f3fc3061 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 17 Jul 2026 10:19:26 +1000 Subject: [PATCH 19/63] feat: add tool, force repo-url, revision and config to be non-nullable in workflows table --- app/db/models/core.py | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/app/db/models/core.py b/app/db/models/core.py index 31ded49..b3d804b 100644 --- a/app/db/models/core.py +++ b/app/db/models/core.py @@ -43,10 +43,11 @@ class Workflow(Base): id: Mapped[UUID] = mapped_column(UUID(as_uuid=True), primary_key=True, default=uuid4) name: Mapped[str] = mapped_column(Text, nullable=False) description: Mapped[str | None] = mapped_column(Text, nullable=True) - repo_url: Mapped[str | None] = mapped_column(Text, nullable=True) - default_revision: Mapped[str | None] = mapped_column(Text, nullable=True) - config_path: Mapped[str | None] = mapped_column(Text, nullable=True) + repo_url: Mapped[str] = mapped_column(Text, nullable=False) + default_revision: Mapped[str] = mapped_column(Text, nullable=False) + config_path: Mapped[str] = mapped_column(Text, nullable=False) prerun_script_path: Mapped[str | None] = mapped_column(Text, nullable=True) + tool: Mapped[str | None] = mapped_column(Text, nullable=True) runs: Mapped[list[WorkflowRun]] = relationship(back_populates="workflow") From bb01dfd342eed0b4cbb185ec49b991a6d069d4ae Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 17 Jul 2026 10:39:33 +1000 Subject: [PATCH 20/63] feat: add migration --- ...nd_require_core_fields_on__7a377a05de6e.py | 41 +++++++++++++++++++ 1 file changed, 41 insertions(+) create mode 100644 alembic/versions/20260717_103718_add_tool_and_require_core_fields_on__7a377a05de6e.py diff --git a/alembic/versions/20260717_103718_add_tool_and_require_core_fields_on__7a377a05de6e.py b/alembic/versions/20260717_103718_add_tool_and_require_core_fields_on__7a377a05de6e.py new file mode 100644 index 0000000..ae06edb --- /dev/null +++ b/alembic/versions/20260717_103718_add_tool_and_require_core_fields_on__7a377a05de6e.py @@ -0,0 +1,41 @@ +"""add_tool_and_require_core_fields_on_workflows""" + +from alembic import op +import sqlalchemy as sa + + +# revision identifiers, used by Alembic. +revision = '7a377a05de6e' +down_revision = '74ab4858580d' +branch_labels = None +depends_on = None + + +def upgrade() -> None: + # ### commands auto generated by Alembic - please adjust! ### + op.add_column('workflows', sa.Column('tool', sa.Text(), nullable=True)) + op.alter_column('workflows', 'repo_url', + existing_type=sa.TEXT(), + nullable=False) + op.alter_column('workflows', 'default_revision', + existing_type=sa.TEXT(), + nullable=False) + op.alter_column('workflows', 'config_path', + existing_type=sa.TEXT(), + nullable=False) + # ### end Alembic commands ### + + +def downgrade() -> None: + # ### commands auto generated by Alembic - please adjust! ### + op.alter_column('workflows', 'config_path', + existing_type=sa.TEXT(), + nullable=True) + op.alter_column('workflows', 'default_revision', + existing_type=sa.TEXT(), + nullable=True) + op.alter_column('workflows', 'repo_url', + existing_type=sa.TEXT(), + nullable=True) + op.drop_column('workflows', 'tool') + # ### end Alembic commands ### From 51361c15be22cbed31e216fe5b7dd17da09c1c63 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 17 Jul 2026 10:40:00 +1000 Subject: [PATCH 21/63] tests: update all workflows in tests --- tests/test_routes_results.py | 42 ++++++-- tests/test_routes_workflow_jobs_extra.py | 9 +- tests/test_routes_workflows.py | 130 ----------------------- tests/test_services_job_utils.py | 37 ++++++- tests/test_services_results_utils.py | 7 +- 5 files changed, 82 insertions(+), 143 deletions(-) diff --git a/tests/test_routes_results.py b/tests/test_routes_results.py index cd63741..18fbf8d 100644 --- a/tests/test_routes_results.py +++ b/tests/test_routes_results.py @@ -24,7 +24,12 @@ def _configure_bindcraft_run(run: WorkflowRun) -> None: - run.workflow = Workflow(name="de-novo-design") + run.workflow = Workflow( + name="de-novo-design", + repo_url="https://github.com/test/de-novo-design", + default_revision="main", + config_path="/config/de-novo-design.config", + ) run.submitted_form_data = {"mode": "bindcraft"} @@ -332,7 +337,12 @@ async def test_get_result_downloads_returns_presigned_links_for_tracked_outputs( name="Results User 5", email="results5@example.com", ) - workflow = Workflow(name="de-novo-design") + workflow = Workflow( + name="de-novo-design", + repo_url="https://github.com/test/de-novo-design", + default_revision="main", + config_path="/config/de-novo-design.config", + ) run = WorkflowRun( owner=user, workflow=workflow, @@ -492,7 +502,12 @@ async def test_get_result_downloads_returns_presigned_links_for_proteinfold_outp name=f"Proteinfold Downloads {tool}", email=f"results-proteinfold-downloads-{tool}@example.com", ) - workflow = Workflow(name="single-prediction") + workflow = Workflow( + name="single-prediction", + repo_url="https://github.com/test/single-prediction", + default_revision="main", + config_path="/config/single-prediction.config", + ) run = WorkflowRun( owner=user, workflow=workflow, @@ -622,7 +637,12 @@ async def test_get_result_snapshots_returns_presigned_links_for_tracked_outputs( name="Results User Snapshots 1", email="results-snapshots1@example.com", ) - workflow = Workflow(name="de-novo-design") + workflow = Workflow( + name="de-novo-design", + repo_url="https://github.com/test/de-novo-design", + default_revision="main", + config_path="/config/de-novo-design.config", + ) run = WorkflowRun( workflow=workflow, owner=user, @@ -973,7 +993,12 @@ async def test_get_result_setting_params_overlays_queued_job_payload(test_db): name="Queued Job User", email="queued-job@example.com", ) - workflow = Workflow(name="de-novo-design-queued") + workflow = Workflow( + name="de-novo-design-queued", + repo_url="https://github.com/test/de-novo-design-queued", + default_revision="main", + config_path="/config/de-novo-design-queued.config", + ) run = WorkflowRun( owner=user, workflow=workflow, @@ -1011,7 +1036,12 @@ async def test_get_result_setting_params_queued_job_invalid_yaml_kept_as_string( name="Queued Job User 2", email="queued-job2@example.com", ) - workflow = Workflow(name="de-novo-design-queued-2") + workflow = Workflow( + name="de-novo-design-queued-2", + repo_url="https://github.com/test/de-novo-design-queued-2", + default_revision="main", + config_path="/config/de-novo-design-queued-2.config", + ) run = WorkflowRun( owner=user, workflow=workflow, diff --git a/tests/test_routes_workflow_jobs_extra.py b/tests/test_routes_workflow_jobs_extra.py index 87d3633..959ae81 100644 --- a/tests/test_routes_workflow_jobs_extra.py +++ b/tests/test_routes_workflow_jobs_extra.py @@ -94,7 +94,14 @@ async def test_get_job_details_success(test_db): name="User", email="user@example.com", ) - workflow = Workflow(id=uuid4(), name="BindCraft", description="Binding workflow") + workflow = Workflow( + id=uuid4(), + name="BindCraft", + description="Binding workflow", + repo_url="https://github.com/test/bindcraft", + default_revision="main", + config_path="/config/bindcraft.config", + ) run = WorkflowRun( id=uuid4(), owner_user_id=user.id, diff --git a/tests/test_routes_workflows.py b/tests/test_routes_workflows.py index 187bef1..2366b3a 100644 --- a/tests/test_routes_workflows.py +++ b/tests/test_routes_workflows.py @@ -477,59 +477,6 @@ def test_extract_final_design_count_string_number(): assert _extract_final_design_count(_form_data(number_of_final_designs="25")) == 25 -# ============================================================================= -# Tests for missing repo_url / default_revision -# ============================================================================= - - -def test_launch_missing_repo_url(client: TestClient, app, test_engine): - """Workflow missing repo_url should return 500.""" - with Session(test_engine) as db: - db.add( - Workflow( - id=uuid4(), - name="single-prediction", - description="No repo workflow", - repo_url=None, - default_revision="dev", - ) - ) - db.commit() - - payload = { - "launch": {"workflow": "single-prediction", "tool": "colabfold", "runName": "test-run"}, - "s3InputKey": "inputs/samplesheets/test.csv", - "formData": {"workflow": "single-prediction", "tool": "colabfold"}, - } - response = client.post("/api/workflows/launch", json=payload) - assert response.status_code == 500 - assert "missing repo_url" in response.json()["detail"] - - -def test_launch_missing_default_revision(client: TestClient, app, test_engine): - """Workflow missing default_revision should return 500.""" - with Session(test_engine) as db: - db.add( - Workflow( - id=uuid4(), - name="single-prediction", - description="No revision workflow", - repo_url="https://github.com/test/norev", - default_revision=None, - ) - ) - db.commit() - - payload = { - "launch": {"workflow": "single-prediction", "tool": "colabfold", "runName": "test-run"}, - "s3InputKey": "inputs/samplesheets/test.csv", - "formData": {"workflow": "single-prediction", "tool": "colabfold"}, - } - response = client.post("/api/workflows/launch", json=payload) - assert response.status_code == 500 - assert "missing default_revision" in response.json()["detail"] - - # ============================================================================= # Tests for proteinfold launch path # ============================================================================= @@ -771,62 +718,6 @@ def _get_db(): yield c -@pytest.fixture -def wisps_no_config_client(test_engine): - """Test client with an interaction-screening workflow that has config_path=None.""" - from sqlalchemy.orm import sessionmaker - - from app.main import create_app - - application = create_app() - user_id = UUID("11111111-1111-1111-1111-111111111111") - - SessionLocal = sessionmaker( - bind=test_engine, autocommit=False, autoflush=False, expire_on_commit=False - ) - setup_session = SessionLocal() - - if not setup_session.get(AppUser, user_id): - setup_session.add( - AppUser( - id=user_id, - auth0_user_id="auth0|test-user", - name="Test User", - email="test@example.com", - ) - ) - - setup_session.add( - Workflow( - id=uuid4(), - name="interaction-screening", - description="WISPS workflow without config_path", - repo_url="https://github.com/test/wisps", - default_revision="main", - config_path=None, - ) - ) - - setup_session.commit() - setup_session.close() - - def _get_db(): - db = SessionLocal() - try: - yield db - finally: - db.close() - - from app.routes.dependencies import require_workflow_execution_role - - application.dependency_overrides[get_db] = _get_db - application.dependency_overrides[get_current_user_id] = lambda: user_id - application.dependency_overrides[require_workflow_execution_role] = lambda: None - - with TestClient(application) as c: - yield c - - @patch("app.routes.workflows.prepare_wisps_workflow", side_effect=_queue_job_for_route_prepare) def test_launch_interaction_screening_success(mock_prepare, wisps_client: TestClient, test_engine): """Test successful interaction-screening workflow launch.""" @@ -994,27 +885,6 @@ def test_launch_interaction_screening_queue_preparation_error( assert count == 0 -def test_launch_interaction_screening_missing_config_path(wisps_no_config_client: TestClient): - """interaction-screening workflow missing config_path should return 500.""" - payload = { - "launch": { - "workflow": "interaction-screening", - "tool": "boltz", - "runName": "wisps-run-no-cfg", - }, - "s3InputKey": "inputs/samplesheets/test.csv", - "formData": { - "workflow": "interaction-screening", - "tool": "boltz", - "fastaS3Uri": "s3://bucket/test.fasta", - "splitOutputDir": "/data/split", - }, - } - - response = wisps_no_config_client.post("/api/workflows/launch", json=payload) - - assert response.status_code == 500 - assert "config_path" in response.json()["detail"] @patch("app.routes.workflows.prepare_wisps_workflow", side_effect=_queue_job_for_route_prepare) diff --git a/tests/test_services_job_utils.py b/tests/test_services_job_utils.py index 4fe5ad6..c4a169d 100644 --- a/tests/test_services_job_utils.py +++ b/tests/test_services_job_utils.py @@ -43,7 +43,12 @@ def commit(self): def _configure_bindcraft_run(run: WorkflowRun) -> None: - run.workflow = Workflow(name="de-novo-design") + run.workflow = Workflow( + name="de-novo-design", + repo_url="https://github.com/test/de-novo-design", + default_revision="main", + config_path="/config/de-novo-design.config", + ) run.tool = "bindcraft" run.submitted_form_data = {"mode": "bindcraft"} @@ -69,8 +74,20 @@ def test_get_user_job_list_rows_returns_current_user_run_metadata(test_db): name="Other User", email="other-jobs@example.com", ) - workflow = Workflow(name="de-novo-design", description="Design workflow") - other_workflow = Workflow(name="other-workflow", description="Other workflow") + workflow = Workflow( + name="de-novo-design", + description="Design workflow", + repo_url="https://github.com/test/de-novo-design", + default_revision="main", + config_path="/config/de-novo-design.config", + ) + other_workflow = Workflow( + name="other-workflow", + description="Other workflow", + repo_url="https://github.com/test/other-workflow", + default_revision="main", + config_path="/config/other-workflow.config", + ) test_db.add_all([user, other_user, workflow, other_workflow]) test_db.commit() @@ -243,7 +260,12 @@ async def test_ensure_completed_run_score_persists_spec_score(test_db): name="Score User", email="score-user@example.com", ) - workflow = Workflow(name="de-novo-design") + workflow = Workflow( + name="de-novo-design", + repo_url="https://github.com/test/de-novo-design", + default_revision="main", + config_path="/config/de-novo-design.config", + ) run = WorkflowRun( owner=user, workflow=workflow, @@ -472,7 +494,12 @@ async def test_get_result_snapshot_downloads_skips_s3_for_proteinfold_workflows( name="Proteinfold No Snapshot User", email="proteinfold-no-snapshot-user@example.com", ) - workflow = Workflow(name="single-prediction") + workflow = Workflow( + name="single-prediction", + repo_url="https://github.com/test/single-prediction", + default_revision="main", + config_path="/config/single-prediction.config", + ) run = WorkflowRun( owner=user, workflow=workflow, diff --git a/tests/test_services_results_utils.py b/tests/test_services_results_utils.py index de2732b..c403544 100644 --- a/tests/test_services_results_utils.py +++ b/tests/test_services_results_utils.py @@ -234,7 +234,12 @@ async def test_get_all_downloads_zipped_writes_category_label_files_and_reads_ea user = AppUserFactory.create_sync() run = WorkflowRunFactory.create_sync( owner=user, - workflow=Workflow(name="de-novo-design"), + workflow=Workflow( + name="de-novo-design", + repo_url="https://github.com/test/de-novo-design", + default_revision="main", + config_path="/config/de-novo-design.config", + ), tool="bindcraft", seqera_run_id="wf-zip-results", ) From af16c330c5f6cc68587d11ac8fa92b45b11df54b Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 17 Jul 2026 10:41:25 +1000 Subject: [PATCH 22/63] feat: update schema diagram --- docs/schema_diagram | 3 + docs/schema_diagram.svg | 422 ++++++++++++++++++++-------------------- 2 files changed, 217 insertions(+), 208 deletions(-) diff --git a/docs/schema_diagram b/docs/schema_diagram index b554556..d078054 100644 --- a/docs/schema_diagram +++ b/docs/schema_diagram @@ -53,6 +53,9 @@ digraph { prerun_script_path TEXT () + + tool + TEXT () > URL="http://Workflow_details.html"] Workflow -> WorkflowRun [label=runs color="#1E88E5" style=dashed tooltip="Relation between Workflow and WorkflowRun"] WorkflowRun [label=< diff --git a/docs/schema_diagram.svg b/docs/schema_diagram.svg index 5f36042..8d030aa 100644 --- a/docs/schema_diagram.svg +++ b/docs/schema_diagram.svg @@ -1,62 +1,62 @@ - - - - + + + AppUser - - -AppUser + + +AppUser + + +id + + +UUID (PK) -id +auth0_user_id -UUID (PK) +TEXT (Unique) -auth0_user_id +name -TEXT (Unique) +TEXT () -name +email -TEXT () +TEXT (Unique) -email +credit -TEXT (Unique) +BIGINT () -credit +credit_updated_at -BIGINT () +DATETIME () -credit_updated_at +credit_updated_by -DATETIME () - - -credit_updated_by - - -TEXT () +TEXT () @@ -64,81 +64,81 @@ WorkflowRun - - -WorkflowRun + + +WorkflowRun + + +id + + +UUID (PK) -id +workflow_id -UUID (PK) +UUID () -workflow_id +owner_user_id -UUID () +UUID () -owner_user_id +seqera_run_id -UUID () +TEXT () -seqera_run_id +binder_name -TEXT () +TEXT () -binder_name +sample_id -TEXT () +TEXT () -sample_id +run_name -TEXT () +TEXT () -run_name +submitted_form_data -TEXT () +JSON () -submitted_form_data +work_dir -JSON () +TEXT () -work_dir +launch_ip -TEXT () +TEXT () -launch_ip +submission_timestamp -TEXT () +DATETIME () -submission_timestamp +tool -DATETIME () - - -tool - - -TEXT () +TEXT () @@ -146,71 +146,77 @@ AppUser->WorkflowRun - - + + -workflow_runs +workflow_runs WorkflowRun->AppUser - - + + -owner +owner Workflow - - -Workflow + + +Workflow + + +id + + +UUID (PK) -id +name -UUID (PK) +TEXT () -name +description -TEXT () +TEXT () -description +repo_url -TEXT () +TEXT () -repo_url +default_revision -TEXT () +TEXT () -default_revision +config_path -TEXT () +TEXT () -config_path +prerun_script_path -TEXT () +TEXT () -prerun_script_path +tool -TEXT () +TEXT () @@ -218,37 +224,37 @@ WorkflowRun->Workflow - - + + -workflow +workflow RunMetric - - -RunMetric - - -run_id - - -UUID (PK) - - -max_score - - -NUMERIC(8, 2) () - - -final_design_count - - -BIGINT () + + +RunMetric + + +run_id + + +UUID (PK) + + +max_score + + +NUMERIC(8, 2) () + + +final_design_count + + +BIGINT () @@ -256,11 +262,11 @@ WorkflowRun->RunMetric - - + + -metrics +metrics @@ -268,19 +274,19 @@ -RunInput +RunInput -run_id +run_id -UUID (PK) +UUID (PK) -s3_object_id +s3_object_id -TEXT (PK) +TEXT (PK) @@ -288,31 +294,31 @@ WorkflowRun->RunInput - + -inputs +inputs RunOutput - - -RunOutput - - -run_id - - -UUID (PK) - - -s3_object_id - - -TEXT (PK) + + +RunOutput + + +run_id + + +UUID (PK) + + +s3_object_id + + +TEXT (PK) @@ -320,41 +326,41 @@ WorkflowRun->RunOutput - - + + -outputs +outputs Workflow->WorkflowRun - - + + -runs +runs RunMetric->WorkflowRun - - + + -run +run RunInput->WorkflowRun - - + + -run +run @@ -362,31 +368,31 @@ -S3Object +S3Object -object_key +object_key -TEXT (PK) +TEXT (PK) -URI +URI -TEXT () +TEXT () -version_id +version_id -TEXT () +TEXT () -size_bytes +size_bytes -BIGINT () +BIGINT () @@ -404,17 +410,17 @@ RunOutput->WorkflowRun - - + + -run +run RunOutput->S3Object - + @@ -434,8 +440,8 @@ S3Object->RunOutput - - + + run_outputs @@ -444,75 +450,75 @@ QueuedJob - - -QueuedJob + + +QueuedJob + + +id + + +UUID (PK) -id +workflow_run_id -UUID (PK) +UUID () -workflow_run_id +workflow_id -UUID () +UUID () -workflow_id +launch_payload -UUID () +JSON () -launch_payload +status -JSON () +VARCHAR(20) () -status +attempts -VARCHAR(20) () +INTEGER () -attempts +queued_at -INTEGER () +DATETIME () -queued_at +last_attempt_at -DATETIME () +DATETIME () -last_attempt_at +next_attempt_at -DATETIME () +DATETIME () -next_attempt_at +submitted_at -DATETIME () +DATETIME () -submitted_at +error -DATETIME () - - -error - - -TEXT () +TEXT () @@ -520,59 +526,59 @@ QueuedJob->WorkflowRun - - + + -workflow_run +workflow_run QueuedJob->Workflow - - + + -workflow +workflow SystemStatusCache - - -SystemStatusCache + + +SystemStatusCache + + +key + + +TEXT (PK) -key +payload -TEXT (PK) +JSON () -payload +checked_at -JSON () +DATETIME () -checked_at +expires_at -DATETIME () +DATETIME () -expires_at +updated_at -DATETIME () - - -updated_at - - -DATETIME () +DATETIME () From 9a57bbabb9b47b16b849f998da0068e652d8c94a Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 17 Jul 2026 10:42:41 +1000 Subject: [PATCH 23/63] chore: lint --- tests/test_routes_workflows.py | 2 -- 1 file changed, 2 deletions(-) diff --git a/tests/test_routes_workflows.py b/tests/test_routes_workflows.py index 2366b3a..b9a339d 100644 --- a/tests/test_routes_workflows.py +++ b/tests/test_routes_workflows.py @@ -885,8 +885,6 @@ def test_launch_interaction_screening_queue_preparation_error( assert count == 0 - - @patch("app.routes.workflows.prepare_wisps_workflow", side_effect=_queue_job_for_route_prepare) def test_launch_with_workflow_field_in_launch(mock_prepare, wisps_client: TestClient, test_engine): """The new frontend format using launch.workflow is accepted alongside launch.tool.""" From 43e5ad28947a252c5888e23d21e3ee79c2523387 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 17 Jul 2026 10:44:26 +1000 Subject: [PATCH 24/63] feat: update tool in admin dashboard --- app/db/admin.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/app/db/admin.py b/app/db/admin.py index 184704d..04fad8a 100644 --- a/app/db/admin.py +++ b/app/db/admin.py @@ -126,6 +126,7 @@ class WorkflowAdmin(ModelView): fields = [ "id", "name", + "tool", "description", "repo_url", "default_revision", @@ -135,6 +136,7 @@ class WorkflowAdmin(ModelView): _NULLABLE_FIELDS = ( "description", + "tool", "prerun_script_path", ) From 6fb4f5757c5b3b48021e4a88ceb4a054815df54c Mon Sep 17 00:00:00 2001 From: Minh Vu Date: Fri, 17 Jul 2026 10:51:00 +1000 Subject: [PATCH 25/63] fix: update output to support cif file formats --- app/services/results_utils.py | 6 +++--- tests/test_services_results_utils.py | 9 +++++++++ 2 files changed, 12 insertions(+), 3 deletions(-) diff --git a/app/services/results_utils.py b/app/services/results_utils.py index 9276a81..f77061b 100644 --- a/app/services/results_utils.py +++ b/app/services/results_utils.py @@ -474,7 +474,7 @@ def classify_boltz_proteinfold_output( sample_id_pattern = re.escape(sample_id) if sample_id else "single_prediction" return classify_proteinfold_output_key( key, - pdb_pattern=rf"/boltz/top_ranked_structures/{sample_id_pattern}\.pdb", + pdb_pattern=rf"/boltz/top_ranked_structures/{sample_id_pattern}\.(?:cif|pdb)", # Find across all subfolders stats_pattern=rf"/boltz/{sample_id_pattern}/.+\.tsv", alignment_pattern=rf"/mmseqs/{sample_id_pattern}\.a3m", @@ -487,7 +487,7 @@ def classify_alphafold2_proteinfold_output( sample_id_pattern = re.escape(sample_id) if sample_id else "single_prediction" return classify_proteinfold_output_key( key, - pdb_pattern=rf"/alphafold2/split_msa_prediction/top_ranked_structures/{sample_id_pattern}\.pdb", + pdb_pattern=rf"/alphafold2/split_msa_prediction/top_ranked_structures/{sample_id_pattern}\.(?:cif|pdb)", stats_pattern=rf"/alphafold2/split_msa_prediction/{sample_id_pattern}/.+\.tsv", ) @@ -498,7 +498,7 @@ def classify_colabfold_proteinfold_output( sample_id_pattern = re.escape(sample_id) if sample_id else "single_prediction" return classify_proteinfold_output_key( key, - pdb_pattern=rf"/colabfold/top_ranked_structures/{sample_id_pattern}\.pdb", + pdb_pattern=rf"/colabfold/top_ranked_structures/{sample_id_pattern}\.(?:cif|pdb)", stats_pattern=rf"/colabfold/{sample_id_pattern}/.+\.tsv", alignment_pattern=rf"/mmseqs/{sample_id_pattern}\.a3m", ) diff --git a/tests/test_services_results_utils.py b/tests/test_services_results_utils.py index de2732b..f1514ca 100644 --- a/tests/test_services_results_utils.py +++ b/tests/test_services_results_utils.py @@ -437,6 +437,9 @@ def test_boltz_proteinfold_helpers_classify_keys_and_build_prefixes(): assert classify_boltz_proteinfold_output( f"{run.id}/boltz/top_ranked_structures/T1024.pdb", "T1024" ) == ClassifiedOutput("pdb", "T1024.pdb") + assert classify_boltz_proteinfold_output( + f"{run.id}/boltz/top_ranked_structures/T1024.cif", "T1024" + ) == ClassifiedOutput("pdb", "T1024.cif") assert classify_boltz_proteinfold_output( f"{run.id}/boltz/T1024/abcd1234.tsv", "T1024" ) == ClassifiedOutput("stats_csv", "abcd1234.tsv") @@ -491,6 +494,9 @@ def test_alphafold2_proteinfold_helpers_classify_keys_and_build_prefixes(): assert classify_alphafold2_proteinfold_output( f"{run.id}/alphafold2/split_msa_prediction/top_ranked_structures/T1024.pdb", "T1024" ) == ClassifiedOutput("pdb", "T1024.pdb") + assert classify_alphafold2_proteinfold_output( + f"{run.id}/alphafold2/split_msa_prediction/top_ranked_structures/T1024.cif", "T1024" + ) == ClassifiedOutput("pdb", "T1024.cif") assert classify_alphafold2_proteinfold_output( f"{run.id}/alphafold2/split_msa_prediction/T1024/abcd1234.tsv", "T1024" ) == ClassifiedOutput("stats_csv", "abcd1234.tsv") @@ -546,6 +552,9 @@ def test_colabfold_proteinfold_helpers_classify_keys_and_build_prefixes(): assert classify_colabfold_proteinfold_output( f"{run.id}/colabfold/top_ranked_structures/T1024.pdb", "T1024" ) == ClassifiedOutput("pdb", "T1024.pdb") + assert classify_colabfold_proteinfold_output( + f"{run.id}/colabfold/top_ranked_structures/T1024.cif", "T1024" + ) == ClassifiedOutput("pdb", "T1024.cif") assert classify_colabfold_proteinfold_output( f"{run.id}/colabfold/T1024/abcd1234.tsv", "T1024" ) == ClassifiedOutput("stats_csv", "abcd1234.tsv") From 0b6442d4b8f9a03218b31ca300c2fae35f017400 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 17 Jul 2026 10:59:48 +1000 Subject: [PATCH 26/63] feat: add workflows name and tool constraint --- ..._constraints_for_workflow__14cff803a3b9.py | 25 +++++++++++++++++++ app/db/models/core.py | 19 ++++++++++++++ 2 files changed, 44 insertions(+) create mode 100644 alembic/versions/20260717_105701_add_unique_constraints_for_workflow__14cff803a3b9.py diff --git a/alembic/versions/20260717_105701_add_unique_constraints_for_workflow__14cff803a3b9.py b/alembic/versions/20260717_105701_add_unique_constraints_for_workflow__14cff803a3b9.py new file mode 100644 index 0000000..0b86f33 --- /dev/null +++ b/alembic/versions/20260717_105701_add_unique_constraints_for_workflow__14cff803a3b9.py @@ -0,0 +1,25 @@ +"""add_unique_constraints_for_workflow_name_tool""" + +from alembic import op +import sqlalchemy as sa + + +# revision identifiers, used by Alembic. +revision = '14cff803a3b9' +down_revision = '7a377a05de6e' +branch_labels = None +depends_on = None + + +def upgrade() -> None: + # ### commands auto generated by Alembic - please adjust! ### + op.create_unique_constraint(op.f('uq_workflows_name'), 'workflows', ['name', 'tool']) + op.create_index('uq_workflows_name_null_tool', 'workflows', ['name'], unique=True, postgresql_where=sa.text('tool IS NULL')) + # ### end Alembic commands ### + + +def downgrade() -> None: + # ### commands auto generated by Alembic - please adjust! ### + op.drop_index('uq_workflows_name_null_tool', table_name='workflows', postgresql_where=sa.text('tool IS NULL')) + op.drop_constraint(op.f('uq_workflows_name'), 'workflows', type_='unique') + # ### end Alembic commands ### diff --git a/app/db/models/core.py b/app/db/models/core.py index b3d804b..8440057 100644 --- a/app/db/models/core.py +++ b/app/db/models/core.py @@ -8,10 +8,12 @@ BigInteger, DateTime, ForeignKey, + Index, Numeric, PrimaryKeyConstraint, Text, UniqueConstraint, + text, ) from sqlalchemy.dialects.postgresql import INET, UUID from sqlalchemy.orm import Mapped, Session, mapped_column, relationship @@ -39,6 +41,23 @@ class AppUser(Base): class Workflow(Base): __tablename__ = "workflows" + __table_args__ = ( + # At most one row per (name, tool) for workflows with tool-specific + # configs (e.g. de-novo-design: bindcraft vs rfdiffusion). + UniqueConstraint("name", "tool"), + # At most one generic (tool-independent) row per name. sqlite_where is + # needed alongside postgresql_where so the tests' SQLite schema (built + # via Base.metadata.create_all) enforces the same partial uniqueness + # as the Postgres migration; without it, SQLite silently drops the + # WHERE clause and the index becomes a full unique constraint on name. + Index( + "uq_workflows_name_null_tool", + "name", + unique=True, + postgresql_where=text("tool IS NULL"), + sqlite_where=text("tool IS NULL"), + ), + ) id: Mapped[UUID] = mapped_column(UUID(as_uuid=True), primary_key=True, default=uuid4) name: Mapped[str] = mapped_column(Text, nullable=False) From 42cea4ddcef131456c95bf94e64b3081a2c83c62 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 17 Jul 2026 11:00:22 +1000 Subject: [PATCH 27/63] feat: query tool if workflows.tool is not null --- app/routes/workflows.py | 21 ++++++++++++++++----- 1 file changed, 16 insertions(+), 5 deletions(-) diff --git a/app/routes/workflows.py b/app/routes/workflows.py index 8500277..77d2ff4 100644 --- a/app/routes/workflows.py +++ b/app/routes/workflows.py @@ -12,7 +12,7 @@ from fastapi import APIRouter, Depends, HTTPException, Query, status from pydantic import ValidationError -from sqlalchemy import CursorResult, func, select, update +from sqlalchemy import CursorResult, func, or_, select, update from sqlalchemy.orm import Session from ..db.models import QueuedJob @@ -207,16 +207,27 @@ async def launch_workflow( ) # Workflow repo_url and revision come from the DB entry for this workflow name - # ("single-prediction", "de-novo-design", etc.). + # ("single-prediction", "de-novo-design", etc.). A workflow's tool column is + # NULL for a single row shared by all of its tools (e.g. single-prediction), + # or set on multiple rows when each tool needs its own repo_url/config_path/ + # default_revision (e.g. de-novo-design: bindcraft vs rfdiffusion). An exact + # tool match is preferred over the generic NULL-tool row when both exist. + tool_matches = func.lower(Workflow.tool) == selected_tool.lower() workflow = db_session.scalar( - select(Workflow).where(func.lower(Workflow.name) == requested_workflow) + select(Workflow) + .where( + func.lower(Workflow.name) == requested_workflow, + or_(Workflow.tool.is_(None), tool_matches), + ) + .order_by(tool_matches.desc()) + .limit(1) ) if not workflow: raise HTTPException( status_code=status.HTTP_500_INTERNAL_SERVER_ERROR, detail=( - f"Workflow '{payload.launch.workflow}' is not configured in workflows table. " - "Seed the workflows catalog before launching." + f"Workflow '{payload.launch.workflow}' with tool '{selected_tool}' is not " + "configured in workflows table. Seed the workflows catalog before launching." ), ) From c1d18c869b5a51cb6f2ea124eeb5987dcd425f56 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 17 Jul 2026 11:00:52 +1000 Subject: [PATCH 28/63] test: add tool in tests --- tests/conftest.py | 1 + tests/scheduler/test_job_queue.py | 2 +- tests/test_routes_workflows.py | 19 +++++++++++++++++++ 3 files changed, 21 insertions(+), 1 deletion(-) diff --git a/tests/conftest.py b/tests/conftest.py index f51b41d..e1f0a0b 100644 --- a/tests/conftest.py +++ b/tests/conftest.py @@ -182,6 +182,7 @@ def app(test_engine): Workflow( id=uuid4(), name="de-novo-design", + tool="bindcraft", description="Test workflow", repo_url="https://github.com/test/repo", default_revision="dev", diff --git a/tests/scheduler/test_job_queue.py b/tests/scheduler/test_job_queue.py index 8b3f7ba..4ed0225 100644 --- a/tests/scheduler/test_job_queue.py +++ b/tests/scheduler/test_job_queue.py @@ -21,7 +21,7 @@ def _get_db(): def _create_queued_job(*, status: str = "pending", next_attempt_at: datetime | None = None): user = AppUserFactory.create_sync() - workflow = WorkflowFactory.create_sync(name="de-novo-design") + workflow = WorkflowFactory.create_sync() workflow_run = WorkflowRunFactory.create_sync( owner=user, workflow=workflow, diff --git a/tests/test_routes_workflows.py b/tests/test_routes_workflows.py index b9a339d..8c75616 100644 --- a/tests/test_routes_workflows.py +++ b/tests/test_routes_workflows.py @@ -60,6 +60,7 @@ def role_check_client(test_engine): Workflow( id=uuid4(), name="de-novo-design", + tool="bindcraft", description="Test workflow", repo_url="https://github.com/test/repo", default_revision="dev", @@ -202,6 +203,24 @@ def test_launch_queue_preparation_error(mock_prepare, client: TestClient, test_e assert count == 0 +def test_launch_de_novo_design_tool_mismatch_returns_500(client: TestClient): + """de-novo-design is matched on tool; an unconfigured tool for it is a 500.""" + payload = { + "launch": { + "workflow": "de-novo-design", + "tool": "rfdiffusion", + "runName": "test-run", + }, + "s3InputKey": "inputs/samplesheets/test.csv", + "formData": {"workflow": "de-novo-design", "tool": "rfdiffusion"}, + } + + response = client.post("/api/workflows/launch", json=payload) + + assert response.status_code == 500 + assert "rfdiffusion" in response.json()["detail"] + + def test_launch_invalid_payload(client: TestClient): """Test launch with invalid payload.""" payload = { From aa81cf6fd8a909157da72cb68d4a747b2a4f59e9 Mon Sep 17 00:00:00 2001 From: Minh Vu Date: Fri, 17 Jul 2026 14:00:29 +1000 Subject: [PATCH 29/63] fix: remove CIF format support from AlphaFold and ColabFold --- app/services/results_utils.py | 4 ++-- tests/test_services_results_utils.py | 6 ------ 2 files changed, 2 insertions(+), 8 deletions(-) diff --git a/app/services/results_utils.py b/app/services/results_utils.py index f77061b..93c24e4 100644 --- a/app/services/results_utils.py +++ b/app/services/results_utils.py @@ -487,7 +487,7 @@ def classify_alphafold2_proteinfold_output( sample_id_pattern = re.escape(sample_id) if sample_id else "single_prediction" return classify_proteinfold_output_key( key, - pdb_pattern=rf"/alphafold2/split_msa_prediction/top_ranked_structures/{sample_id_pattern}\.(?:cif|pdb)", + pdb_pattern=rf"/alphafold2/split_msa_prediction/top_ranked_structures/{sample_id_pattern}\.pdb", stats_pattern=rf"/alphafold2/split_msa_prediction/{sample_id_pattern}/.+\.tsv", ) @@ -498,7 +498,7 @@ def classify_colabfold_proteinfold_output( sample_id_pattern = re.escape(sample_id) if sample_id else "single_prediction" return classify_proteinfold_output_key( key, - pdb_pattern=rf"/colabfold/top_ranked_structures/{sample_id_pattern}\.(?:cif|pdb)", + pdb_pattern=rf"/colabfold/top_ranked_structures/{sample_id_pattern}\.pdb", stats_pattern=rf"/colabfold/{sample_id_pattern}/.+\.tsv", alignment_pattern=rf"/mmseqs/{sample_id_pattern}\.a3m", ) diff --git a/tests/test_services_results_utils.py b/tests/test_services_results_utils.py index f1514ca..95b5e79 100644 --- a/tests/test_services_results_utils.py +++ b/tests/test_services_results_utils.py @@ -494,9 +494,6 @@ def test_alphafold2_proteinfold_helpers_classify_keys_and_build_prefixes(): assert classify_alphafold2_proteinfold_output( f"{run.id}/alphafold2/split_msa_prediction/top_ranked_structures/T1024.pdb", "T1024" ) == ClassifiedOutput("pdb", "T1024.pdb") - assert classify_alphafold2_proteinfold_output( - f"{run.id}/alphafold2/split_msa_prediction/top_ranked_structures/T1024.cif", "T1024" - ) == ClassifiedOutput("pdb", "T1024.cif") assert classify_alphafold2_proteinfold_output( f"{run.id}/alphafold2/split_msa_prediction/T1024/abcd1234.tsv", "T1024" ) == ClassifiedOutput("stats_csv", "abcd1234.tsv") @@ -552,9 +549,6 @@ def test_colabfold_proteinfold_helpers_classify_keys_and_build_prefixes(): assert classify_colabfold_proteinfold_output( f"{run.id}/colabfold/top_ranked_structures/T1024.pdb", "T1024" ) == ClassifiedOutput("pdb", "T1024.pdb") - assert classify_colabfold_proteinfold_output( - f"{run.id}/colabfold/top_ranked_structures/T1024.cif", "T1024" - ) == ClassifiedOutput("pdb", "T1024.cif") assert classify_colabfold_proteinfold_output( f"{run.id}/colabfold/T1024/abcd1234.tsv", "T1024" ) == ClassifiedOutput("stats_csv", "abcd1234.tsv") From 359988dafb842ee64fd3c66a87760f94ed46624c Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 17 Jul 2026 15:45:40 +1000 Subject: [PATCH 30/63] fix: drop stale name-only unique index before workflows tool migration --- ..._constraints_for_workflow__14cff803a3b9.py | 2 +- ...name_only_unique_index_on__7616ae331fe5.py | 24 +++++++++++++++++++ 2 files changed, 25 insertions(+), 1 deletion(-) create mode 100644 alembic/versions/20260717_153902_drop_stale_name_only_unique_index_on__7616ae331fe5.py diff --git a/alembic/versions/20260717_105701_add_unique_constraints_for_workflow__14cff803a3b9.py b/alembic/versions/20260717_105701_add_unique_constraints_for_workflow__14cff803a3b9.py index 0b86f33..5f05e41 100644 --- a/alembic/versions/20260717_105701_add_unique_constraints_for_workflow__14cff803a3b9.py +++ b/alembic/versions/20260717_105701_add_unique_constraints_for_workflow__14cff803a3b9.py @@ -6,7 +6,7 @@ # revision identifiers, used by Alembic. revision = '14cff803a3b9' -down_revision = '7a377a05de6e' +down_revision = '7616ae331fe5' branch_labels = None depends_on = None diff --git a/alembic/versions/20260717_153902_drop_stale_name_only_unique_index_on__7616ae331fe5.py b/alembic/versions/20260717_153902_drop_stale_name_only_unique_index_on__7616ae331fe5.py new file mode 100644 index 0000000..86872cb --- /dev/null +++ b/alembic/versions/20260717_153902_drop_stale_name_only_unique_index_on__7616ae331fe5.py @@ -0,0 +1,24 @@ +"""drop_stale_name_only_unique_index_on_workflows""" + +from alembic import op +import sqlalchemy as sa + + +# revision identifiers, used by Alembic. +revision = '7616ae331fe5' +down_revision = '7a377a05de6e' +branch_labels = None +depends_on = None + + +def upgrade() -> None: + # Some environments have a plain unique index named 'uq_workflows_name' on + # workflows.name alone, predating this feature and untracked by any prior + # migration (likely created out-of-band). It collides with the + # same-named (name, tool) constraint the next migration creates. + op.execute("DROP INDEX IF EXISTS uq_workflows_name") + + +def downgrade() -> None: + # The dropped index was untracked drift, not a schema feature; nothing to restore. + pass From ccf74897421bc1c415ff76891acec13b3167771c Mon Sep 17 00:00:00 2001 From: laclac102 <53896516+vtnphan@users.noreply.github.com> Date: Tue, 21 Jul 2026 10:27:52 +1000 Subject: [PATCH 31/63] feat: proteindj config and executor --- app/services/proteindj_config.py | 56 +++++++++++ app/services/proteindj_executor.py | 155 +++++++++++++++++++++++++++++ 2 files changed, 211 insertions(+) create mode 100644 app/services/proteindj_config.py create mode 100644 app/services/proteindj_executor.py diff --git a/app/services/proteindj_config.py b/app/services/proteindj_config.py new file mode 100644 index 0000000..832e58e --- /dev/null +++ b/app/services/proteindj_config.py @@ -0,0 +1,56 @@ +"""ProteinDJ workflow configuration and executor settings (modeled after bindflow).""" + +from __future__ import annotations + +from typing import Any + +from ..schemas.workflows import WorkflowUserDetails +from .cluster_utils import encode_ip +from .workflow_config_fetcher import fetch_workflow_config + + +def get_proteindj_default_params( + out_dir: str, + input_pdb: str | None = None, + hotspot_residues: str | None = None, + num_designs: int | None = None, + design_length: str | None = None, +) -> dict[str, Any]: + """Get default parameters for proteindj workflow. + + ProteinDJ (rfdiffusion) takes a single PDB plus design params directly — + no samplesheet — so these are passed straight through as paramsText keys. + """ + params: dict[str, Any] = {"outdir": out_dir} + if input_pdb is not None: + params["input_pdb"] = input_pdb + if hotspot_residues is not None: + params["hotspot_residues"] = hotspot_residues + if num_designs is not None: + params["num_designs"] = num_designs + if design_length is not None: + params["design_length"] = design_length + return params + + +def get_proteindj_config_profiles() -> list[str]: + """Get config profiles for proteindj workflow.""" + return ["singularity"] + + +def get_proteindj_config_text( + config_file_path: str, + *, + user_details: WorkflowUserDetails, +) -> str: + """Read proteindj base config and append a process override block with runtime values.""" + base = fetch_workflow_config(config_file_path) + + account = ( + f"{user_details.user_email}:{encode_ip(user_details.ip_address)}" + if user_details.ip_address + else user_details.user_email + ) + cluster_opts = f"-A {account}" + override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' + return base + override diff --git a/app/services/proteindj_executor.py b/app/services/proteindj_executor.py new file mode 100644 index 0000000..58182d9 --- /dev/null +++ b/app/services/proteindj_executor.py @@ -0,0 +1,155 @@ +"""ProteinDJ workflow executor for Seqera Platform (modeled after bindflow).""" + +from __future__ import annotations + +import logging +import os +from datetime import UTC, datetime +from typing import Any + +from sqlalchemy.orm import Session + +from ..db.models import QueuedJob, WorkflowRun +from ..schemas.workflows import WorkflowFormData, WorkflowLaunchForm, WorkflowUserDetails +from .launch_payloads import get_executor_script, inject_prerun_script, without_prerun_script +from .proteindj_config import ( + get_proteindj_config_profiles, + get_proteindj_config_text, + get_proteindj_default_params, +) +from .seqera import ( + WorkflowLaunchResult, + _get_required_env, + params_to_yaml_text, + post_seqera_launch, +) +from .seqera_errors import SeqeraConfigurationError + +logger = logging.getLogger(__name__) + + +def _design_length(form_data: WorkflowFormData) -> str | None: + # The frontend's Input Configuration step (shared with bindcraft) sends + # min_length/max_length as separate fields; ProteinDJ expects them as a + # single "min-max" range. + extra = form_data.extra_fields + min_length = extra.get("min_length") + max_length = extra.get("max_length") + if min_length is None or max_length is None: + return None + return f"{min_length}-{max_length}" + + +async def prepare_proteindj_workflow( # pylint: disable=too-many-locals + form: WorkflowLaunchForm, + *, + db_session: Session, + workflow_run: WorkflowRun, + pipeline: str, + config_path: str, + revision: str | None = None, + output_id: str | None = None, + form_data: WorkflowFormData, + user_details: WorkflowUserDetails, + commit: bool = False, +) -> QueuedJob: + """Build and queue a proteindj launch payload.""" + workspace_id = _get_required_env("WORK_SPACE") + compute_env_id = _get_required_env("COMPUTE_ID") + work_dir = _get_required_env("WORK_DIR") + s3_bucket = _get_required_env("AWS_S3_BUCKET") + + run_name = (form.runName or "").strip() + if not run_name: + raise SeqeraConfigurationError("Missing run name for workflow launch") + # Always use a unique backend-generated ID for outputs to avoid S3 prefix collisions. + output_key = (output_id or "").strip() + if not output_key: + raise SeqeraConfigurationError("Missing output identifier for workflow launch") + out_dir = f"s3://{s3_bucket}/{output_key}" + + extra = form_data.extra_fields + default_params = get_proteindj_default_params( + out_dir, + input_pdb=extra.get("starting_pdb"), + hotspot_residues=extra.get("target_hotspot_residues"), + num_designs=extra.get("number_of_final_designs"), + design_length=_design_length(form_data), + ) + + # Serialize to YAML + params_text = params_to_yaml_text(default_params) + + # Add custom paramsText from frontend if provided + if form.paramsText and form.paramsText.strip(): + params_text = f"{params_text}\n{form.paramsText.rstrip()}" + + launch_payload: dict[str, Any] = { + "computeEnvId": compute_env_id, + "runName": run_name, + "pipeline": pipeline, + "workDir": work_dir, + "workspaceId": workspace_id, + "revision": revision or "dev", + "paramsText": params_text, + "configProfiles": get_proteindj_config_profiles(), + "configText": get_proteindj_config_text( + config_path, + user_details=user_details, + ), + "resume": False, + } + + queued_job = QueuedJob( + workflow=workflow_run.workflow, + workflow_run=workflow_run, + launch_payload=without_prerun_script(launch_payload), + status="pending", + next_attempt_at=datetime.now(UTC), + ) + db_session.add(queued_job) + if commit: + db_session.commit() + else: + db_session.flush() + return queued_job + + +async def launch_proteindj_workflow( # pylint: disable=too-many-locals + *, + queued_job: QueuedJob, + dry_run: bool = False, +) -> WorkflowLaunchResult | None: + """Launch a proteindj workflow on the Seqera Platform.""" + launch_payload = queued_job.launch_payload + + # Log the complete params being sent + logger.info("Launch payload paramsText", extra={"paramsText": launch_payload["paramsText"]}) + + logger.info( + "Launching proteindj workflow via Seqera API", + extra={ + "workspaceId": launch_payload["workspaceId"], + "computeEnvId": launch_payload["computeEnvId"], + "pipeline": launch_payload["pipeline"], + "runName": launch_payload["runName"], + }, + ) + + prerun_script = get_executor_script( + prerun_script_path=queued_job.workflow.prerun_script_path, + module_loads=["singularity", "nextflow"], + env={ + "AWS_ACCESS_KEY_ID": os.getenv("AWS_ACCESS_KEY_ID", ""), + "AWS_SECRET_ACCESS_KEY": os.getenv("AWS_SECRET_ACCESS_KEY", ""), + "AWS_REGION": os.getenv("AWS_REGION", "ap-southeast-2"), + }, + ) + runtime_payload = inject_prerun_script( + launch_payload=launch_payload, prerun_script=prerun_script + ) + + if dry_run: + logger.info("Dry run - not launching proteindj workflow") + return None + return await post_seqera_launch({"launch": runtime_payload}, workflow_label="ProteinDJ") From 82604b5e2a9742f29ff4a86ad37f98202abd439e Mon Sep 17 00:00:00 2001 From: laclac102 <53896516+vtnphan@users.noreply.github.com> Date: Tue, 21 Jul 2026 10:54:42 +1000 Subject: [PATCH 32/63] feat: wire proteindj workflow to de-novo-design --- app/routes/workflows.py | 48 ++++++++++++++++++++++++++--------------- app/scheduler/jobs.py | 8 ++++++- 2 files changed, 38 insertions(+), 18 deletions(-) diff --git a/app/routes/workflows.py b/app/routes/workflows.py index 77d2ff4..89ba515 100644 --- a/app/routes/workflows.py +++ b/app/routes/workflows.py @@ -44,6 +44,7 @@ upload_csv_to_s3, upload_wisps_samplesheet_to_s3, ) +from ..services.proteindj_executor import prepare_proteindj_workflow from ..services.proteinfold_executor import prepare_proteinfold_workflow from ..services.s3 import S3ConfigurationError, S3ServiceError, generate_presigned_url from ..services.seqera_errors import SeqeraConfigurationError @@ -131,7 +132,7 @@ def _extract_final_design_count(form_data: WorkflowFormData | None) -> int | Non return None try: parsed = int(str(value).strip()) - except TypeError, ValueError: + except (TypeError, ValueError): return None return parsed if parsed >= 1 else None @@ -350,23 +351,36 @@ async def launch_workflow( user_details=user_details, ) elif workflow_name in ("de-novo-design", "bindflow", "bindcraft"): - # de-novo-design → bindflow executor. - # selected_tool carries the chosen algorithm ("bindcraft", "rfdiffusion"). + # de-novo-design → bindflow executor (bindcraft) or proteindj executor + # (rfdiffusion), depending on the chosen algorithm. tool_mode = selected_tool - bindcraft_launch_form = payload.launch.model_copy(update={"runName": seqera_run_name}) - queued_job = await prepare_bindflow_workflow( - bindcraft_launch_form, - s3_input_key, - db_session=db_session, - workflow_run=workflow_run, - pipeline=workflow.repo_url, - config_path=workflow.config_path, - revision=workflow.default_revision, - output_id=str(run_id), - mode=tool_mode, - form_data=payload.formData, - user_details=user_details, - ) + de_novo_launch_form = payload.launch.model_copy(update={"runName": seqera_run_name}) + if tool_mode.lower() == "rfdiffusion": + queued_job = await prepare_proteindj_workflow( + de_novo_launch_form, + db_session=db_session, + workflow_run=workflow_run, + pipeline=workflow.repo_url, + config_path=workflow.config_path, + revision=workflow.default_revision, + output_id=str(run_id), + form_data=payload.formData, + user_details=user_details, + ) + else: + queued_job = await prepare_bindflow_workflow( + de_novo_launch_form, + s3_input_key, + db_session=db_session, + workflow_run=workflow_run, + pipeline=workflow.repo_url, + config_path=workflow.config_path, + revision=workflow.default_revision, + output_id=str(run_id), + mode=tool_mode, + form_data=payload.formData, + user_details=user_details, + ) elif workflow_name in ("interaction-screening", "bulk-prediction"): assert wisps_form_data is not None wisps_launch_form = payload.launch.model_copy(update={"runName": seqera_run_name}) diff --git a/app/scheduler/jobs.py b/app/scheduler/jobs.py index 177821c..3ce5ec2 100644 --- a/app/scheduler/jobs.py +++ b/app/scheduler/jobs.py @@ -14,6 +14,7 @@ from ..schemas.workflows import WorkflowName from ..services import health, seqera from ..services.bindflow_executor import launch_bindflow_workflow +from ..services.proteindj_executor import launch_proteindj_workflow from ..services.proteinfold_executor import launch_proteinfold_workflow from ..services.seqera import WorkflowLaunchResult from ..services.seqera_errors import SeqeraAPIError, SeqeraConfigurationError @@ -75,7 +76,12 @@ def launch_job(job_id: UUID, dry_run: bool = False) -> None: elif workflow_name in ("single-prediction", "proteinfold"): launch_func = launch_proteinfold_workflow elif workflow_name in ("de-novo-design", "bindflow", "bindcraft"): - launch_func = launch_bindflow_workflow + # de-novo-design covers two algorithms (bindcraft vs rfdiffusion), each + # with its own executor; workflow_run.tool holds the one selected at launch. + tool = (job.workflow_run.tool or "").lower() + launch_func = ( + launch_proteindj_workflow if tool == "rfdiffusion" else launch_bindflow_workflow + ) else: raise ValueError(f"Unsupported workflow: {job.workflow.name}") try: From 0b7a2a68abf1788deeaaef0c8ab1ecce3cb25a1d Mon Sep 17 00:00:00 2001 From: Minh Vu Date: Tue, 21 Jul 2026 14:13:27 +1000 Subject: [PATCH 33/63] feat: add single-prediction entity validation and related tests --- app/routes/workflows.py | 42 +++++++++++ app/schemas/workflows.py | 78 ++++++++++++++++++++ tests/test_routes_workflows.py | 126 ++++++++++++++++++++++++++++++++- tests/test_schemas.py | 70 ++++++++++++++++++ 4 files changed, 313 insertions(+), 3 deletions(-) diff --git a/app/routes/workflows.py b/app/routes/workflows.py index 8500277..068cab2 100644 --- a/app/routes/workflows.py +++ b/app/routes/workflows.py @@ -24,12 +24,14 @@ ListRunsResponse, RunInputPresignedUrlResponse, S3DatasetUploadResponse, + SinglePredictionEntity, WispsDatasetUploadRequest, WispsFormData, WorkflowFormData, WorkflowLaunchPayload, WorkflowLaunchResponse, WorkflowUserDetails, + validate_single_prediction_entities, ) from ..services.bindflow_executor import _get_required_env, prepare_bindflow_workflow from ..services.credits import ( @@ -136,6 +138,43 @@ def _extract_final_design_count(form_data: WorkflowFormData | None) -> int | Non return parsed if parsed >= 1 else None +def _validate_single_prediction_form(form_data: WorkflowFormData, tool: str) -> None: + """Validate single-prediction entity limits; raise HTTPException on failure. + + Guards against jobs with too many entities or has no protein input + """ + raw_entities = form_data.extra_fields.get("entities") + if not isinstance(raw_entities, list): + raise HTTPException( + status_code=status.HTTP_422_UNPROCESSABLE_ENTITY, + detail="'entities' is required in formData for single-prediction.", + ) + try: + entities = [SinglePredictionEntity.model_validate(item) for item in raw_entities] + except ValidationError as exc: + raise HTTPException( + status_code=status.HTTP_422_UNPROCESSABLE_ENTITY, + detail="Invalid entity data in formData for single-prediction.", + ) from exc + + raw_potentials = form_data.extra_fields.get("boltz_use_potentials") + boltz_use_potentials = ( + raw_potentials.strip().lower() in ("true", "1", "yes") + if isinstance(raw_potentials, str) + else bool(raw_potentials) + ) + + try: + validate_single_prediction_entities( + entities, tool, boltz_use_potentials=boltz_use_potentials + ) + except ValueError as exc: + raise HTTPException( + status_code=status.HTTP_422_UNPROCESSABLE_ENTITY, + detail=str(exc), + ) from exc + + @router.post("/me/sync") async def sync_current_user( current_user_id: UUID = Depends(get_current_user_id), @@ -317,6 +356,9 @@ async def launch_workflow( detail=f"'{missing}' is required in formData for {workflow_name}.", ) from exc + if workflow_name in ("single-prediction", "proteinfold"): + _validate_single_prediction_form(payload.formData, selected_tool) + try: queued_job: QueuedJob seqera_run_name = build_unique_run_name(payload.launch.runName or "") diff --git a/app/schemas/workflows.py b/app/schemas/workflows.py index a19370d..a549761 100644 --- a/app/schemas/workflows.py +++ b/app/schemas/workflows.py @@ -2,6 +2,7 @@ from __future__ import annotations +import re from datetime import datetime from enum import StrEnum from typing import Any, Literal @@ -13,6 +14,15 @@ ] WorkflowTool = Literal["alphafold2", "bindcraft", "boltz", "colabfold", "rfdiffusion"] +SINGLE_PREDICTION_MAX_ENTITIES = 52 +SINGLE_PREDICTION_LIGAND_SIZE = 30 +SINGLE_PREDICTION_SIZE_LIMITS: dict[str, int] = { + "alphafold2": 2000, + "colabfold": 4000, + "boltz": 4000, +} +SINGLE_PREDICTION_BOLTZ_POTENTIALS_SIZE_LIMIT = 2000 + class PipelineStatus(StrEnum): """Pipeline status values from Seqera Platform.""" @@ -111,6 +121,74 @@ class WispsFormData(WorkflowFormData): ) +class SinglePredictionEntity(BaseModel): + """A single entity row submitted for a single-prediction run.""" + + model_config = ConfigDict(extra="allow") + + id: str | None = None + moleculeType: str + copyNumber: int = 1 + sequence: str = "" + + +def _entity_prediction_size(entity: SinglePredictionEntity) -> int: + """Prediction size for one copy of an entity (ligands use a fixed size).""" + if entity.moleculeType in ("ligand", "ccd"): + return SINGLE_PREDICTION_LIGAND_SIZE + return len(re.sub(r"\s+", "", entity.sequence or "")) + + +def single_prediction_size_limit(tool: str, boltz_use_potentials: bool) -> int: + """Exclusive upper bound on the total prediction size for the given tool.""" + if tool == "boltz" and boltz_use_potentials: + return SINGLE_PREDICTION_BOLTZ_POTENTIALS_SIZE_LIMIT + return SINGLE_PREDICTION_SIZE_LIMITS.get(tool, SINGLE_PREDICTION_SIZE_LIMITS["colabfold"]) + + +def validate_single_prediction_entities( + entities: list[SinglePredictionEntity], + tool: str, + *, + boltz_use_potentials: bool = False, +) -> None: + """Validate entity count, protein presence, and total prediction size. + + Raises ``ValueError`` with a user-facing message on the first violation. + """ + if not entities: + raise ValueError("At least one entity is required.") + + total_copies = 0 + total_size = 0 + has_protein = False + for entity in entities: + if entity.copyNumber < 1: + raise ValueError("Each entity must have a copy number of at least 1.") + total_copies += entity.copyNumber + if entity.moleculeType == "protein": + has_protein = True + total_size += _entity_prediction_size(entity) * entity.copyNumber + + if total_copies > SINGLE_PREDICTION_MAX_ENTITIES: + raise ValueError( + f"Too many entities: {total_copies} including copies. " + f"The maximum allowed is {SINGLE_PREDICTION_MAX_ENTITIES}." + ) + + if not has_protein: + raise ValueError( + "At least one entity must be a protein. " + "This workflow cannot run without a protein input." + ) + + limit = single_prediction_size_limit(tool, boltz_use_potentials) + if total_size >= limit: + raise ValueError( + f"Total prediction size ({total_size}) must be less than {limit} for {tool}." + ) + + class WorkflowLaunchPayload(BaseModel): model_config = ConfigDict(extra="forbid") diff --git a/tests/test_routes_workflows.py b/tests/test_routes_workflows.py index 187bef1..12e09c1 100644 --- a/tests/test_routes_workflows.py +++ b/tests/test_routes_workflows.py @@ -564,7 +564,18 @@ def test_launch_proteinfold_success(mock_prepare, client: TestClient, test_engin payload = { "launch": {"workflow": "single-prediction", "tool": "colabfold", "runName": "pf-run-1"}, "s3InputKey": "inputs/samplesheets/test.csv", - "formData": {"workflow": "single-prediction", "tool": "colabfold"}, + "formData": { + "workflow": "single-prediction", + "tool": "colabfold", + "entities": [ + { + "id": "seq1", + "moleculeType": "protein", + "copyNumber": 1, + "sequence": "ACDEFGHIK", + } + ], + }, } response = client.post("/api/workflows/launch", json=payload) @@ -600,7 +611,18 @@ def test_launch_proteinfold_queue_preparation_configuration_error( "runName": "pf-run-cfg-err", }, "s3InputKey": "inputs/samplesheets/test.csv", - "formData": {"workflow": "single-prediction", "tool": "colabfold"}, + "formData": { + "workflow": "single-prediction", + "tool": "colabfold", + "entities": [ + { + "id": "seq1", + "moleculeType": "protein", + "copyNumber": 1, + "sequence": "ACDEFGHIK", + } + ], + }, } response = client.post("/api/workflows/launch", json=payload) @@ -621,7 +643,18 @@ def test_launch_proteinfold_queue_preparation_error(mock_prepare, client: TestCl "runName": "pf-run-exec-err", }, "s3InputKey": "inputs/samplesheets/test.csv", - "formData": {"workflow": "single-prediction", "tool": "colabfold"}, + "formData": { + "workflow": "single-prediction", + "tool": "colabfold", + "entities": [ + { + "id": "seq1", + "moleculeType": "protein", + "copyNumber": 1, + "sequence": "ACDEFGHIK", + } + ], + }, } response = client.post("/api/workflows/launch", json=payload) @@ -629,6 +662,93 @@ def test_launch_proteinfold_queue_preparation_error(mock_prepare, client: TestCl assert response.json()["detail"] == "Failed to queue local workflow run." +# ============================================================================= +# Tests for single-prediction entity validation +# ============================================================================= + + +def _single_prediction_payload(entities, tool="colabfold", **form_extra): + form_data = {"workflow": "single-prediction", "tool": tool, "entities": entities} + form_data.update(form_extra) + return { + "launch": {"workflow": "single-prediction", "tool": tool, "runName": "sp-val"}, + "s3InputKey": "inputs/samplesheets/test.csv", + "formData": form_data, + } + + +def _protein_entity(sequence="ACDEFGHIK", copy_number=1): + return { + "id": "seq1", + "moleculeType": "protein", + "copyNumber": copy_number, + "sequence": sequence, + } + + +def test_launch_single_prediction_rejects_missing_entities(client: TestClient, test_engine): + _add_proteinfold_workflow(test_engine) + payload = { + "launch": {"workflow": "single-prediction", "tool": "colabfold", "runName": "sp-no-ent"}, + "s3InputKey": "inputs/samplesheets/test.csv", + "formData": {"workflow": "single-prediction", "tool": "colabfold"}, + } + response = client.post("/api/workflows/launch", json=payload) + assert response.status_code == 422 + assert "entities" in response.json()["detail"] + + +def test_launch_single_prediction_rejects_too_many_entities(client: TestClient, test_engine): + _add_proteinfold_workflow(test_engine) + payload = _single_prediction_payload([_protein_entity(copy_number=53)]) + response = client.post("/api/workflows/launch", json=payload) + assert response.status_code == 422 + assert "Too many entities" in response.json()["detail"] + + +def test_launch_single_prediction_requires_protein(client: TestClient, test_engine): + _add_proteinfold_workflow(test_engine) + entities = [ + {"id": "seq1", "moleculeType": "dna", "copyNumber": 1, "sequence": "ACGT"}, + ] + payload = _single_prediction_payload(entities, tool="boltz") + response = client.post("/api/workflows/launch", json=payload) + assert response.status_code == 422 + assert "must be a protein" in response.json()["detail"] + + +def test_launch_single_prediction_rejects_oversized_alphafold2(client: TestClient, test_engine): + _add_proteinfold_workflow(test_engine) + payload = _single_prediction_payload( + [_protein_entity(sequence="A" * 2000)], tool="alphafold2" + ) + response = client.post("/api/workflows/launch", json=payload) + assert response.status_code == 422 + assert "less than 2000" in response.json()["detail"] + + +@patch( + "app.routes.workflows.prepare_proteinfold_workflow", side_effect=_queue_job_for_route_prepare +) +def test_launch_single_prediction_boltz_potentials_reduces_limit( + mock_prepare, client: TestClient, test_engine +): + _add_proteinfold_workflow(test_engine) + + ok_payload = _single_prediction_payload( + [_protein_entity(sequence="A" * 1999)], tool="boltz", boltz_use_potentials=True + ) + assert client.post("/api/workflows/launch", json=ok_payload).status_code == 201 + + over_payload = _single_prediction_payload( + [_protein_entity(sequence="A" * 2000)], tool="boltz", boltz_use_potentials=True + ) + response = client.post("/api/workflows/launch", json=over_payload) + assert response.status_code == 422 + assert "less than 2000" in response.json()["detail"] + mock_prepare.assert_called_once() + + # ============================================================================= # Tests for require_workflow_execution_role # ============================================================================= diff --git a/tests/test_schemas.py b/tests/test_schemas.py index fbe5200..895e9ad 100644 --- a/tests/test_schemas.py +++ b/tests/test_schemas.py @@ -17,6 +17,7 @@ ListRunsResponse, PipelineStatus, RunInfo, + SinglePredictionEntity, UIStatus, WispsDatasetUploadRequest, WispsSequenceItem, @@ -24,6 +25,8 @@ WorkflowLaunchPayload, WorkflowLaunchResponse, map_pipeline_status_to_ui, + single_prediction_size_limit, + validate_single_prediction_entities, ) @@ -462,3 +465,70 @@ def test_interaction_screening_request_extra_fields_forbidden(): runId="run-1", extra="bad", ) + + +# ============================================================================= +# Single-prediction entity validation +# ============================================================================= + + +def _protein(sequence="ACDEFGHIK", copy_number=1): + return SinglePredictionEntity( + id="seq1", moleculeType="protein", copyNumber=copy_number, sequence=sequence + ) + + +def test_size_limit_defaults_and_boltz_potentials(): + assert single_prediction_size_limit("colabfold", False) == 4000 + assert single_prediction_size_limit("boltz", False) == 4000 + assert single_prediction_size_limit("alphafold2", False) == 2000 + assert single_prediction_size_limit("boltz", True) == 2000 + assert single_prediction_size_limit("colabfold", True) == 4000 + + +def test_validate_single_prediction_accepts_valid_input(): + validate_single_prediction_entities([_protein()], "colabfold") + + +def test_validate_single_prediction_requires_entities(): + with pytest.raises(ValueError, match="At least one entity"): + validate_single_prediction_entities([], "colabfold") + + +def test_validate_single_prediction_counts_copies_toward_limit(): + validate_single_prediction_entities([_protein(copy_number=52)], "colabfold") + with pytest.raises(ValueError, match="Too many entities"): + validate_single_prediction_entities([_protein(copy_number=53)], "colabfold") + + +def test_validate_single_prediction_requires_protein(): + dna = SinglePredictionEntity(id="d", moleculeType="dna", copyNumber=1, sequence="ACGT") + with pytest.raises(ValueError, match="must be a protein"): + validate_single_prediction_entities([dna], "boltz") + + +def test_validate_single_prediction_rejects_copy_number_below_one(): + entity = SinglePredictionEntity(id="p", moleculeType="protein", copyNumber=0, sequence="ACD") + with pytest.raises(ValueError, match="copy number of at least 1"): + validate_single_prediction_entities([entity], "colabfold") + + +def test_validate_single_prediction_ligand_uses_fixed_size(): + protein = _protein(sequence="A" * 3960) + ligand = SinglePredictionEntity( + id="lig", moleculeType="ligand", copyNumber=1, sequence="CC(=O)O" + ) + # 3960 + 30 (ligand) = 3990 < 4000 + validate_single_prediction_entities([protein, ligand], "colabfold") + # Two ligand copies: 3960 + 60 = 4020 >= 4000 + ligand_two = SinglePredictionEntity( + id="lig", moleculeType="ligand", copyNumber=2, sequence="CC(=O)O" + ) + with pytest.raises(ValueError, match="less than 4000"): + validate_single_prediction_entities([protein, ligand_two], "colabfold") + + +def test_validate_single_prediction_size_limit_is_exclusive(): + with pytest.raises(ValueError, match="less than 2000"): + validate_single_prediction_entities([_protein(sequence="A" * 2000)], "alphafold2") + validate_single_prediction_entities([_protein(sequence="A" * 1999)], "alphafold2") From 1e3ccd023d5d7d185a9cb31dc88657417b446511 Mon Sep 17 00:00:00 2001 From: laclac102 <53896516+vtnphan@users.noreply.github.com> Date: Tue, 21 Jul 2026 14:14:12 +1000 Subject: [PATCH 34/63] feat: add rfd_length params --- app/services/proteindj_config.py | 3 +++ app/services/proteindj_executor.py | 4 +++- 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/app/services/proteindj_config.py b/app/services/proteindj_config.py index 832e58e..c8361d4 100644 --- a/app/services/proteindj_config.py +++ b/app/services/proteindj_config.py @@ -15,6 +15,7 @@ def get_proteindj_default_params( hotspot_residues: str | None = None, num_designs: int | None = None, design_length: str | None = None, + rfd_length: str | None = None, ) -> dict[str, Any]: """Get default parameters for proteindj workflow. @@ -30,6 +31,8 @@ def get_proteindj_default_params( params["num_designs"] = num_designs if design_length is not None: params["design_length"] = design_length + if rfd_length is not None: + params["rfd_length"] = rfd_length return params diff --git a/app/services/proteindj_executor.py b/app/services/proteindj_executor.py index 58182d9..ccdb24b 100644 --- a/app/services/proteindj_executor.py +++ b/app/services/proteindj_executor.py @@ -69,12 +69,14 @@ async def prepare_proteindj_workflow( # pylint: disable=too-many-locals out_dir = f"s3://{s3_bucket}/{output_key}" extra = form_data.extra_fields + length_range = _design_length(form_data) default_params = get_proteindj_default_params( out_dir, input_pdb=extra.get("starting_pdb"), hotspot_residues=extra.get("target_hotspot_residues"), num_designs=extra.get("number_of_final_designs"), - design_length=_design_length(form_data), + design_length=length_range, + rfd_length=length_range, ) # Serialize to YAML From eb227653973cf4768c288c2a84b9f3ef23bb24bd Mon Sep 17 00:00:00 2001 From: laclac102 <53896516+vtnphan@users.noreply.github.com> Date: Tue, 21 Jul 2026 15:16:14 +1000 Subject: [PATCH 35/63] feat: remove afd_length, update related tests --- app/routes/workflows.py | 2 +- app/services/proteindj_config.py | 3 - app/services/proteindj_executor.py | 4 +- tests/scheduler/test_scheduler_jobs.py | 55 ++- tests/test_proteindj_coverage.py | 445 +++++++++++++++++++++++++ tests/test_routes_workflows.py | 73 ++++ 6 files changed, 574 insertions(+), 8 deletions(-) create mode 100644 tests/test_proteindj_coverage.py diff --git a/app/routes/workflows.py b/app/routes/workflows.py index 89ba515..d07feee 100644 --- a/app/routes/workflows.py +++ b/app/routes/workflows.py @@ -132,7 +132,7 @@ def _extract_final_design_count(form_data: WorkflowFormData | None) -> int | Non return None try: parsed = int(str(value).strip()) - except (TypeError, ValueError): + except TypeError, ValueError: return None return parsed if parsed >= 1 else None diff --git a/app/services/proteindj_config.py b/app/services/proteindj_config.py index c8361d4..832e58e 100644 --- a/app/services/proteindj_config.py +++ b/app/services/proteindj_config.py @@ -15,7 +15,6 @@ def get_proteindj_default_params( hotspot_residues: str | None = None, num_designs: int | None = None, design_length: str | None = None, - rfd_length: str | None = None, ) -> dict[str, Any]: """Get default parameters for proteindj workflow. @@ -31,8 +30,6 @@ def get_proteindj_default_params( params["num_designs"] = num_designs if design_length is not None: params["design_length"] = design_length - if rfd_length is not None: - params["rfd_length"] = rfd_length return params diff --git a/app/services/proteindj_executor.py b/app/services/proteindj_executor.py index ccdb24b..58182d9 100644 --- a/app/services/proteindj_executor.py +++ b/app/services/proteindj_executor.py @@ -69,14 +69,12 @@ async def prepare_proteindj_workflow( # pylint: disable=too-many-locals out_dir = f"s3://{s3_bucket}/{output_key}" extra = form_data.extra_fields - length_range = _design_length(form_data) default_params = get_proteindj_default_params( out_dir, input_pdb=extra.get("starting_pdb"), hotspot_residues=extra.get("target_hotspot_residues"), num_designs=extra.get("number_of_final_designs"), - design_length=length_range, - rfd_length=length_range, + design_length=_design_length(form_data), ) # Serialize to YAML diff --git a/tests/scheduler/test_scheduler_jobs.py b/tests/scheduler/test_scheduler_jobs.py index c6747b0..ed800e3 100644 --- a/tests/scheduler/test_scheduler_jobs.py +++ b/tests/scheduler/test_scheduler_jobs.py @@ -20,7 +20,9 @@ async def _failing_launch(**_kwargs): raise RuntimeError("Seqera launch failed") -def _create_queued_job(*, attempts: int = 0, workflow_name: str = "de-novo-design"): +def _create_queued_job( + *, attempts: int = 0, workflow_name: str = "de-novo-design", tool: str | None = None +): user = AppUserFactory.create_sync() workflow = WorkflowFactory.create_sync(name=workflow_name) workflow_run = WorkflowRunFactory.create_sync( @@ -28,6 +30,7 @@ def _create_queued_job(*, attempts: int = 0, workflow_name: str = "de-novo-desig workflow=workflow, work_dir=f"/work/{workflow_name}-{attempts}-{uuid4()}", seqera_run_id=None, + tool=tool, ) return QueuedJobFactory.create_sync( workflow_run=workflow_run, @@ -125,6 +128,56 @@ async def _successful_launch(**kwargs): assert queued_job.next_attempt_at is not None +def test_launch_job_dispatches_proteindj_for_rfdiffusion_tool( + test_db, persistent_models, monkeypatch +): + queued_job = _create_queued_job(tool="rfdiffusion") + calls = [] + + async def _successful_launch(**kwargs): + calls.append(kwargs) + return WorkflowLaunchResult(workflow_id="seqera-run-rfd", status="submitted") + + async def _unexpected_bindflow_launch(**_kwargs): + raise AssertionError("launch_bindflow_workflow should not be called for rfdiffusion") + + monkeypatch.setattr(scheduler_jobs, "get_db", _get_db_override(test_db)) + monkeypatch.setattr(scheduler_jobs, "is_seqera_available", lambda _db: True) + monkeypatch.setattr(scheduler_jobs, "launch_proteindj_workflow", _successful_launch) + monkeypatch.setattr(scheduler_jobs, "launch_bindflow_workflow", _unexpected_bindflow_launch) + + scheduler_jobs.launch_job(queued_job.id) + + test_db.refresh(queued_job) + assert calls == [{"queued_job": queued_job, "dry_run": False}] + assert queued_job.workflow_run.seqera_run_id == "seqera-run-rfd" + assert queued_job.status == "submitted" + + +def test_launch_job_dispatches_bindflow_for_bindcraft_tool(test_db, persistent_models, monkeypatch): + queued_job = _create_queued_job(tool="bindcraft") + calls = [] + + async def _successful_launch(**kwargs): + calls.append(kwargs) + return WorkflowLaunchResult(workflow_id="seqera-run-bc", status="submitted") + + async def _unexpected_proteindj_launch(**_kwargs): + raise AssertionError("launch_proteindj_workflow should not be called for bindcraft") + + monkeypatch.setattr(scheduler_jobs, "get_db", _get_db_override(test_db)) + monkeypatch.setattr(scheduler_jobs, "is_seqera_available", lambda _db: True) + monkeypatch.setattr(scheduler_jobs, "launch_bindflow_workflow", _successful_launch) + monkeypatch.setattr(scheduler_jobs, "launch_proteindj_workflow", _unexpected_proteindj_launch) + + scheduler_jobs.launch_job(queued_job.id) + + test_db.refresh(queued_job) + assert calls == [{"queued_job": queued_job, "dry_run": False}] + assert queued_job.workflow_run.seqera_run_id == "seqera-run-bc" + assert queued_job.status == "submitted" + + def test_launch_job_dispatches_wisps_workflows(test_db, persistent_models, monkeypatch): queued_job = _create_queued_job(workflow_name="interaction-screening") calls = [] diff --git a/tests/test_proteindj_coverage.py b/tests/test_proteindj_coverage.py new file mode 100644 index 0000000..3830748 --- /dev/null +++ b/tests/test_proteindj_coverage.py @@ -0,0 +1,445 @@ +"""Tests to boost coverage for proteindj executor and config modules.""" + +from __future__ import annotations + +from contextlib import contextmanager +from unittest.mock import AsyncMock, Mock, mock_open, patch + +import pytest +from sqlalchemy import select + +from app.db.models import QueuedJob +from app.schemas.workflows import WorkflowFormData, WorkflowLaunchForm, WorkflowUserDetails +from app.services.proteindj_config import ( + get_proteindj_config_profiles, + get_proteindj_config_text, + get_proteindj_default_params, +) +from app.services.proteindj_executor import ( + _design_length, + launch_proteindj_workflow, + prepare_proteindj_workflow, +) +from app.services.seqera import WorkflowLaunchResult +from app.services.seqera_errors import SeqeraConfigurationError +from tests.datagen import AppUserFactory, QueuedJobFactory, WorkflowFactory, WorkflowRunFactory + +_USER_DETAILS = WorkflowUserDetails( + user_email="user@ex.com", + ip_address="1.2.3.4", +) + + +def _form_data(**extra) -> WorkflowFormData: + return WorkflowFormData(workflow="de-novo-design", tool="rfdiffusion", **extra) + + +def _make_launch_form(**kwargs) -> WorkflowLaunchForm: + defaults = { + "workflow": "de-novo-design", + "tool": "rfdiffusion", + "runName": "test-run", + "paramsText": None, + } + defaults.update(kwargs) + return WorkflowLaunchForm(**defaults) + + +def _queued_proteindj_job( + *, + params_text: str | None = None, + prerun_script_path: str | None = None, +) -> QueuedJob: + user = AppUserFactory.create_sync() + workflow = WorkflowFactory.create_sync( + name="de-novo-design", + prerun_script_path=prerun_script_path, + ) + workflow_run = WorkflowRunFactory.create_sync(workflow=workflow, owner=user) + launch_payload = { + "computeEnvId": "ce_456", + "runName": "test-run", + "pipeline": "https://github.com/org/proteindj", + "workDir": "/work/dir", + "workspaceId": "ws_123", + "revision": "dev", + "paramsText": params_text + or ("outdir: s3://my-bucket/run-output-id\ninput_pdb: s3://my-bucket/inputs/test.pdb"), + "configProfiles": ["singularity"], + "configText": "config_text", + "resume": False, + } + return QueuedJobFactory.create_sync( + workflow=workflow, + workflow_run=workflow_run, + launch_payload=launch_payload, + status="pending", + ) + + +@contextmanager +def _mock_proteindj_db_context(): + workflow = Mock(name="workflow") + workflow_run = Mock(name="workflow_run") + workflow_run.workflow = workflow + db_session = Mock(name="db_session") + queued_job = Mock(name="queued_job") + with patch( + "app.services.proteindj_executor.QueuedJob", return_value=queued_job + ) as queued_job_cls: + yield db_session, workflow_run, workflow, queued_job_cls, queued_job + + +@pytest.fixture +def seqera_env(monkeypatch): + """Set required Seqera environment variables for launch tests.""" + monkeypatch.setenv("SEQERA_API_URL", "https://api.seqera.test") + monkeypatch.setenv("SEQERA_ACCESS_TOKEN", "test_token") + monkeypatch.setenv("WORK_SPACE", "ws_123") + monkeypatch.setenv("COMPUTE_ID", "ce_456") + monkeypatch.setenv("WORK_DIR", "/work/dir") + monkeypatch.setenv("AWS_S3_BUCKET", "my-bucket") + monkeypatch.setenv("AWS_ACCESS_KEY_ID", "test_key") + monkeypatch.setenv("AWS_SECRET_ACCESS_KEY", "test_secret") + + +# ============================================================================= +# Tests for get_proteindj_default_params() +# ============================================================================= + + +def test_get_proteindj_default_params_only_outdir(): + params = get_proteindj_default_params("s3://bucket/out") + assert params == {"outdir": "s3://bucket/out"} + + +def test_get_proteindj_default_params_all_fields(): + params = get_proteindj_default_params( + "s3://bucket/out", + input_pdb="s3://bucket/in.pdb", + hotspot_residues="A20,A21", + num_designs=5, + design_length="100-150", + ) + assert params == { + "outdir": "s3://bucket/out", + "input_pdb": "s3://bucket/in.pdb", + "hotspot_residues": "A20,A21", + "num_designs": 5, + "design_length": "100-150", + } + + +def test_get_proteindj_default_params_partial_fields(): + params = get_proteindj_default_params("s3://bucket/out", num_designs=3) + assert params == {"outdir": "s3://bucket/out", "num_designs": 3} + + +# ============================================================================= +# Tests for get_proteindj_config_profiles() +# ============================================================================= + + +def test_get_proteindj_config_profiles_returns_list(): + profiles = get_proteindj_config_profiles() + assert isinstance(profiles, list) + + +def test_get_proteindj_config_profiles_contains_singularity(): + assert "singularity" in get_proteindj_config_profiles() + + +# ============================================================================= +# Tests for get_proteindj_config_text() +# ============================================================================= + + +def test_get_proteindj_config_text_appends_process_block(): + with patch("builtins.open", mock_open(read_data="base_config")): + result = get_proteindj_config_text("/fake/proteindj.config", user_details=_USER_DETAILS) + assert "process {" in result + assert "clusterOptions" in result + + +def test_get_proteindj_config_text_contains_email_and_encoded_ip(): + with patch("builtins.open", mock_open(read_data="base_config")): + result = get_proteindj_config_text("/fake/proteindj.config", user_details=_USER_DETAILS) + assert "user@ex.com" in result + assert "MS4yLjMuNA==" in result + + +def test_get_proteindj_config_text_without_ip_uses_email_only(): + with patch("builtins.open", mock_open(read_data="base_config")): + result = get_proteindj_config_text( + "/fake/proteindj.config", + user_details=_USER_DETAILS.model_copy(update={"ip_address": ""}), + ) + assert "-A user@ex.com" in result + assert ":" not in result.split("clusterOptions = ")[1] + + +def test_get_proteindj_config_text_contains_base_config(): + with patch("builtins.open", mock_open(read_data="base_config")): + result = get_proteindj_config_text("/fake/proteindj.config", user_details=_USER_DETAILS) + assert "base_config" in result + + +# ============================================================================= +# Tests for _design_length() +# ============================================================================= + + +def test_design_length_both_present(): + assert _design_length(_form_data(min_length=100, max_length=150)) == "100-150" + + +def test_design_length_missing_min(): + assert _design_length(_form_data(max_length=150)) is None + + +def test_design_length_missing_max(): + assert _design_length(_form_data(min_length=100)) is None + + +def test_design_length_missing_both(): + assert _design_length(_form_data()) is None + + +# ============================================================================= +# Tests for prepare_proteindj_workflow() +# ============================================================================= + + +@pytest.mark.anyio +async def test_prepare_proteindj_workflow_writes_expected_queued_job( + test_db, persistent_models, seqera_env +): + user = AppUserFactory.create_sync() + workflow = WorkflowFactory.create_sync(name="de-novo-design") + workflow_run = WorkflowRunFactory.create_sync(workflow=workflow, owner=user) + + form = _make_launch_form(runName="queued-proteindj-run") + form_data = _form_data( + starting_pdb="s3://my-bucket/inputs/test.pdb", + target_hotspot_residues="A20,A21", + number_of_final_designs=5, + min_length=100, + max_length=150, + ) + + with ( + patch( + "app.services.proteindj_executor.get_proteindj_config_text", + return_value="config_text", + ), + patch( + "app.services.proteindj_executor.get_proteindj_config_profiles", + return_value=["singularity"], + ), + ): + prepared_job = await prepare_proteindj_workflow( + form=form, + db_session=test_db, + workflow_run=workflow_run, + pipeline="https://github.com/org/proteindj", + config_path="/fake/proteindj.config", + revision="main", + output_id="run-output-id", + form_data=form_data, + user_details=_USER_DETAILS.model_copy( + update={"user_email": "test@example.com", "ip_address": "127.0.0.1"} + ), + ) + + queued_job = test_db.scalar( + select(QueuedJob).where(QueuedJob.workflow_run_id == workflow_run.id) + ) + assert queued_job is not None + assert queued_job.workflow_id == workflow.id + assert queued_job.workflow_run_id == workflow_run.id + assert queued_job.status == "pending" + assert queued_job.next_attempt_at is not None + assert queued_job.id == prepared_job.id + assert queued_job.launch_payload["computeEnvId"] == "ce_456" + assert queued_job.launch_payload["runName"] == "queued-proteindj-run" + assert queued_job.launch_payload["pipeline"] == "https://github.com/org/proteindj" + assert queued_job.launch_payload["workDir"] == "/work/dir" + assert queued_job.launch_payload["workspaceId"] == "ws_123" + assert queued_job.launch_payload["revision"] == "main" + assert queued_job.launch_payload["configProfiles"] == ["singularity"] + assert queued_job.launch_payload["configText"] == "config_text" + assert "preRunScript" not in queued_job.launch_payload + assert queued_job.launch_payload["resume"] is False + params_text = queued_job.launch_payload["paramsText"] + assert "outdir: s3://my-bucket/run-output-id" in params_text + assert "input_pdb: s3://my-bucket/inputs/test.pdb" in params_text + assert "hotspot_residues: A20,A21" in params_text + assert "num_designs: 5" in params_text + assert "design_length: 100-150" in params_text + + +@pytest.mark.anyio +async def test_prepare_proteindj_workflow_appends_custom_params_text( + test_db, persistent_models, seqera_env +): + user = AppUserFactory.create_sync() + workflow = WorkflowFactory.create_sync(name="de-novo-design") + workflow_run = WorkflowRunFactory.create_sync(workflow=workflow, owner=user) + + form = _make_launch_form(paramsText="extra_param: value") + + with ( + patch("app.services.proteindj_executor.get_proteindj_config_text", return_value=""), + patch( + "app.services.proteindj_executor.get_proteindj_config_profiles", + return_value=["singularity"], + ), + ): + prepared_job = await prepare_proteindj_workflow( + form=form, + db_session=test_db, + workflow_run=workflow_run, + pipeline="https://github.com/org/proteindj", + config_path="/fake/proteindj.config", + output_id="run-output-id", + form_data=_form_data(), + user_details=_USER_DETAILS, + ) + + queued_job = test_db.scalar( + select(QueuedJob).where(QueuedJob.workflow_run_id == workflow_run.id) + ) + assert "extra_param: value" in queued_job.launch_payload["paramsText"] + assert prepared_job.id == queued_job.id + + +@pytest.mark.anyio +async def test_prepare_proteindj_workflow_missing_run_name(seqera_env): + form = _make_launch_form(runName=" ") + with ( + _mock_proteindj_db_context() as (db_session, workflow_run, *_), + pytest.raises(SeqeraConfigurationError, match="run name"), + ): + await prepare_proteindj_workflow( + form=form, + db_session=db_session, + workflow_run=workflow_run, + pipeline="https://github.com/org/proteindj", + config_path="/fake/proteindj.config", + output_id="run-output-id", + form_data=_form_data(), + user_details=_USER_DETAILS, + ) + + +@pytest.mark.anyio +async def test_prepare_proteindj_workflow_missing_output_id(seqera_env): + form = _make_launch_form() + with ( + _mock_proteindj_db_context() as (db_session, workflow_run, *_), + pytest.raises(SeqeraConfigurationError, match="output identifier"), + ): + await prepare_proteindj_workflow( + form=form, + db_session=db_session, + workflow_run=workflow_run, + pipeline="https://github.com/org/proteindj", + config_path="/fake/proteindj.config", + output_id=None, + form_data=_form_data(), + user_details=_USER_DETAILS, + ) + + +@pytest.mark.anyio +async def test_prepare_proteindj_workflow_empty_output_id(seqera_env): + form = _make_launch_form() + with ( + _mock_proteindj_db_context() as (db_session, workflow_run, *_), + pytest.raises(SeqeraConfigurationError, match="output identifier"), + ): + await prepare_proteindj_workflow( + form=form, + db_session=db_session, + workflow_run=workflow_run, + pipeline="https://github.com/org/proteindj", + config_path="/fake/proteindj.config", + output_id=" ", + form_data=_form_data(), + user_details=_USER_DETAILS, + ) + + +# ============================================================================= +# Tests for launch_proteindj_workflow() +# ============================================================================= + + +@pytest.mark.anyio +async def test_launch_proteindj_workflow_success(seqera_env, persistent_models): + expected_result = WorkflowLaunchResult( + workflow_id="wf_success", status="submitted", message=None + ) + + with ( + patch( + "app.services.proteindj_executor.post_seqera_launch", + new_callable=AsyncMock, + return_value=expected_result, + ) as mock_post, + ): + result = await launch_proteindj_workflow(queued_job=_queued_proteindj_job()) + + assert result.workflow_id == "wf_success" + assert result.status == "submitted" + mock_post.assert_called_once() + posted_payload = mock_post.call_args.args[0]["launch"] + assert "module load singularity" in posted_payload["preRunScript"] + assert "module load nextflow" in posted_payload["preRunScript"] + assert "export AWS_ACCESS_KEY_ID" in posted_payload["preRunScript"] + + +@pytest.mark.anyio +async def test_launch_proteindj_workflow_with_prerun_script_path(seqera_env, persistent_models): + expected_result = WorkflowLaunchResult(workflow_id="wf_prerun", status="submitted") + + with ( + patch( + "app.services.proteindj_executor.post_seqera_launch", + new_callable=AsyncMock, + return_value=expected_result, + ) as mock_post, + patch( + "app.services.proteindj_executor.get_executor_script", + return_value="prerun_body", + ) as mock_script, + ): + result = await launch_proteindj_workflow( + queued_job=_queued_proteindj_job(prerun_script_path="/some/prerun.sh") + ) + + assert result.workflow_id == "wf_prerun" + posted_payload = mock_post.call_args.args[0]["launch"] + assert posted_payload["preRunScript"] == "prerun_body" + assert mock_script.call_args.kwargs["prerun_script_path"] == "/some/prerun.sh" + + +@pytest.mark.anyio +async def test_launch_proteindj_workflow_dry_run(seqera_env, persistent_models): + with patch( + "app.services.proteindj_executor.post_seqera_launch", + new_callable=AsyncMock, + ) as mock_post: + result = await launch_proteindj_workflow(queued_job=_queued_proteindj_job(), dry_run=True) + + assert result is None + mock_post.assert_not_called() + + +@pytest.mark.anyio +async def test_launch_proteindj_workflow_missing_env_var(monkeypatch, persistent_models): + monkeypatch.delenv("SEQERA_API_URL", raising=False) + monkeypatch.delenv("SEQERA_ACCESS_TOKEN", raising=False) + + with pytest.raises(SeqeraConfigurationError, match="SEQERA_API_URL"): + await launch_proteindj_workflow(queued_job=_queued_proteindj_job()) diff --git a/tests/test_routes_workflows.py b/tests/test_routes_workflows.py index 8c75616..50f55d4 100644 --- a/tests/test_routes_workflows.py +++ b/tests/test_routes_workflows.py @@ -33,6 +33,22 @@ async def _queue_job_for_route_prepare(form, _s3_input_key, **kwargs): return queued_job +async def _queue_job_for_proteindj_route_prepare(form, **kwargs): + # proteindj has no samplesheet, so prepare_proteindj_workflow takes no + # s3_input_key positional arg (unlike prepare_bindflow_workflow above). + db_session = kwargs["db_session"] + workflow_run = kwargs["workflow_run"] + queued_job = QueuedJob( + workflow=workflow_run.workflow, + workflow_run=workflow_run, + launch_payload={"runName": form.runName}, + status="pending", + ) + db_session.add(queued_job) + db_session.flush() + return queued_job + + @pytest.fixture def role_check_client(test_engine): """Test client with auth bypassed but require_workflow_execution_role active.""" @@ -221,6 +237,63 @@ def test_launch_de_novo_design_tool_mismatch_returns_500(client: TestClient): assert "rfdiffusion" in response.json()["detail"] +def _add_rfdiffusion_workflow(test_engine): + """Helper to add a de-novo-design/rfdiffusion workflow row to the test DB.""" + with Session(test_engine) as db: + existing = db.scalar( + select(Workflow).where( + Workflow.name == "de-novo-design", Workflow.tool == "rfdiffusion" + ) + ) + if not existing: + db.add( + Workflow( + id=uuid4(), + name="de-novo-design", + tool="rfdiffusion", + description="ProteinDJ workflow", + repo_url="https://github.com/test/proteindj", + default_revision="dev", + config_path="/some/proteindj.config", + ) + ) + db.commit() + + +@patch("app.routes.workflows.prepare_bindflow_workflow") +@patch( + "app.routes.workflows.prepare_proteindj_workflow", + side_effect=_queue_job_for_proteindj_route_prepare, +) +def test_launch_de_novo_design_rfdiffusion_routes_to_proteindj( + mock_prepare_proteindj, mock_prepare_bindflow, client: TestClient, test_engine +): + """tool='rfdiffusion' on de-novo-design must dispatch to the proteindj executor.""" + _add_rfdiffusion_workflow(test_engine) + + payload = { + "launch": { + "workflow": "de-novo-design", + "tool": "rfdiffusion", + "runName": "rfd-run-1", + }, + "s3InputKey": "inputs/pdb/target.pdb", + "formData": {"workflow": "de-novo-design", "tool": "rfdiffusion"}, + } + + response = client.post("/api/workflows/launch", json=payload) + + assert response.status_code == 201 + data = response.json() + assert data["status"] == "pending" + mock_prepare_proteindj.assert_called_once() + mock_prepare_bindflow.assert_not_called() + assert ( + mock_prepare_proteindj.call_args.kwargs["pipeline"] == "https://github.com/test/proteindj" + ) + assert mock_prepare_proteindj.call_args.kwargs["output_id"] == data["runId"] + + def test_launch_invalid_payload(client: TestClient): """Test launch with invalid payload.""" payload = { From 569162464db6cd2306e8a22a21924d6f5aa7935f Mon Sep 17 00:00:00 2001 From: laclac102 <53896516+vtnphan@users.noreply.github.com> Date: Tue, 21 Jul 2026 16:27:49 +1000 Subject: [PATCH 36/63] fix: out_dir params --- app/services/proteindj_config.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/app/services/proteindj_config.py b/app/services/proteindj_config.py index 832e58e..1161a8b 100644 --- a/app/services/proteindj_config.py +++ b/app/services/proteindj_config.py @@ -21,7 +21,7 @@ def get_proteindj_default_params( ProteinDJ (rfdiffusion) takes a single PDB plus design params directly — no samplesheet — so these are passed straight through as paramsText keys. """ - params: dict[str, Any] = {"outdir": out_dir} + params: dict[str, Any] = {"out_dir": out_dir} if input_pdb is not None: params["input_pdb"] = input_pdb if hotspot_residues is not None: From 2def2a2478b12f74a5e6acb22582f2eaaa96ea97 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Wed, 22 Jul 2026 09:22:06 +1000 Subject: [PATCH 37/63] fix: out_dir params --- app/services/proteindj_config.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/app/services/proteindj_config.py b/app/services/proteindj_config.py index 1161a8b..832e58e 100644 --- a/app/services/proteindj_config.py +++ b/app/services/proteindj_config.py @@ -21,7 +21,7 @@ def get_proteindj_default_params( ProteinDJ (rfdiffusion) takes a single PDB plus design params directly — no samplesheet — so these are passed straight through as paramsText keys. """ - params: dict[str, Any] = {"out_dir": out_dir} + params: dict[str, Any] = {"outdir": out_dir} if input_pdb is not None: params["input_pdb"] = input_pdb if hotspot_residues is not None: From c989abac0e59a4d59305195fe68520dc91c31a2c Mon Sep 17 00:00:00 2001 From: Minh Vu Date: Wed, 22 Jul 2026 09:55:22 +1000 Subject: [PATCH 38/63] fix: black --- tests/test_routes_workflows.py | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/tests/test_routes_workflows.py b/tests/test_routes_workflows.py index 12e09c1..ec55158 100644 --- a/tests/test_routes_workflows.py +++ b/tests/test_routes_workflows.py @@ -719,9 +719,7 @@ def test_launch_single_prediction_requires_protein(client: TestClient, test_engi def test_launch_single_prediction_rejects_oversized_alphafold2(client: TestClient, test_engine): _add_proteinfold_workflow(test_engine) - payload = _single_prediction_payload( - [_protein_entity(sequence="A" * 2000)], tool="alphafold2" - ) + payload = _single_prediction_payload([_protein_entity(sequence="A" * 2000)], tool="alphafold2") response = client.post("/api/workflows/launch", json=payload) assert response.status_code == 422 assert "less than 2000" in response.json()["detail"] From 8a6881d84197455cbcd8827924f62341e27dfbca Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Wed, 22 Jul 2026 10:04:24 +1000 Subject: [PATCH 39/63] feat: add ProteinDjFromData --- app/schemas/workflows.py | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/app/schemas/workflows.py b/app/schemas/workflows.py index a19370d..3b76b0b 100644 --- a/app/schemas/workflows.py +++ b/app/schemas/workflows.py @@ -111,6 +111,26 @@ class WispsFormData(WorkflowFormData): ) +class ProteinDjFormData(WorkflowFormData): + """Form data for the ProteinDJ (rfdiffusion) de-novo-design workflow.""" + + starting_pdb: str = Field(..., description="S3 URI of the uploaded starting PDB file") + target_hotspot_residues: str = Field( + ..., description="Comma-separated hotspot residues, e.g. 'A20,A21'" + ) + number_of_final_designs: int = Field(..., ge=1, description="Number of designs to generate") + min_length: int = Field(..., description="Minimum binder length") + max_length: int = Field(..., description="Maximum binder length") + + @field_validator("starting_pdb", "target_hotspot_residues") + @classmethod + def _not_blank(cls, value: str) -> str: + stripped = value.strip() + if not stripped: + raise ValueError("This field is required") + return stripped + + class WorkflowLaunchPayload(BaseModel): model_config = ConfigDict(extra="forbid") From 76c6f1bec7e7674b40393eeeeb0069629b8488e2 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Wed, 22 Jul 2026 10:05:40 +1000 Subject: [PATCH 40/63] feat: add required fields to ProteinDJ form data --- app/services/proteindj_config.py | 25 +++-- app/services/proteindj_executor.py | 40 +++++--- tests/test_proteindj_coverage.py | 146 ++++++++++++++++++++++++----- 3 files changed, 160 insertions(+), 51 deletions(-) diff --git a/app/services/proteindj_config.py b/app/services/proteindj_config.py index 832e58e..bc83991 100644 --- a/app/services/proteindj_config.py +++ b/app/services/proteindj_config.py @@ -11,26 +11,23 @@ def get_proteindj_default_params( out_dir: str, - input_pdb: str | None = None, - hotspot_residues: str | None = None, - num_designs: int | None = None, - design_length: str | None = None, + input_pdb: str, + hotspot_residues: str, + num_designs: int, + design_length: str, ) -> dict[str, Any]: """Get default parameters for proteindj workflow. ProteinDJ (rfdiffusion) takes a single PDB plus design params directly — no samplesheet — so these are passed straight through as paramsText keys. """ - params: dict[str, Any] = {"outdir": out_dir} - if input_pdb is not None: - params["input_pdb"] = input_pdb - if hotspot_residues is not None: - params["hotspot_residues"] = hotspot_residues - if num_designs is not None: - params["num_designs"] = num_designs - if design_length is not None: - params["design_length"] = design_length - return params + return { + "outdir": out_dir, + "input_pdb": input_pdb, + "hotspot_residues": hotspot_residues, + "num_designs": num_designs, + "design_length": design_length, + } def get_proteindj_config_profiles() -> list[str]: diff --git a/app/services/proteindj_executor.py b/app/services/proteindj_executor.py index 58182d9..0a199ee 100644 --- a/app/services/proteindj_executor.py +++ b/app/services/proteindj_executor.py @@ -7,10 +7,16 @@ from datetime import UTC, datetime from typing import Any +from pydantic import ValidationError from sqlalchemy.orm import Session from ..db.models import QueuedJob, WorkflowRun -from ..schemas.workflows import WorkflowFormData, WorkflowLaunchForm, WorkflowUserDetails +from ..schemas.workflows import ( + ProteinDjFormData, + WorkflowFormData, + WorkflowLaunchForm, + WorkflowUserDetails, +) from .launch_payloads import get_executor_script, inject_prerun_script, without_prerun_script from .proteindj_config import ( get_proteindj_config_profiles, @@ -28,16 +34,24 @@ logger = logging.getLogger(__name__) -def _design_length(form_data: WorkflowFormData) -> str | None: +def _design_length(fields: ProteinDjFormData) -> str: # The frontend's Input Configuration step (shared with bindcraft) sends # min_length/max_length as separate fields; ProteinDJ expects them as a # single "min-max" range. - extra = form_data.extra_fields - min_length = extra.get("min_length") - max_length = extra.get("max_length") - if min_length is None or max_length is None: - return None - return f"{min_length}-{max_length}" + return f"{fields.min_length}-{fields.max_length}" + + +def _parse_proteindj_form_data(form_data: WorkflowFormData) -> ProteinDjFormData: + try: + return ProteinDjFormData.model_validate(form_data.model_dump()) + except ValidationError as exc: + missing = next( + (str(e["loc"][-1]) for e in exc.errors() if e.get("loc")), + "formData", + ) + raise SeqeraConfigurationError( + f"'{missing}' is required in formData for ProteinDJ workflow launch" + ) from exc async def prepare_proteindj_workflow( # pylint: disable=too-many-locals @@ -68,13 +82,13 @@ async def prepare_proteindj_workflow( # pylint: disable=too-many-locals raise SeqeraConfigurationError("Missing output identifier for workflow launch") out_dir = f"s3://{s3_bucket}/{output_key}" - extra = form_data.extra_fields + proteindj_fields = _parse_proteindj_form_data(form_data) default_params = get_proteindj_default_params( out_dir, - input_pdb=extra.get("starting_pdb"), - hotspot_residues=extra.get("target_hotspot_residues"), - num_designs=extra.get("number_of_final_designs"), - design_length=_design_length(form_data), + input_pdb=proteindj_fields.starting_pdb, + hotspot_residues=proteindj_fields.target_hotspot_residues, + num_designs=proteindj_fields.number_of_final_designs, + design_length=_design_length(proteindj_fields), ) # Serialize to YAML diff --git a/tests/test_proteindj_coverage.py b/tests/test_proteindj_coverage.py index 3830748..6204a47 100644 --- a/tests/test_proteindj_coverage.py +++ b/tests/test_proteindj_coverage.py @@ -9,7 +9,12 @@ from sqlalchemy import select from app.db.models import QueuedJob -from app.schemas.workflows import WorkflowFormData, WorkflowLaunchForm, WorkflowUserDetails +from app.schemas.workflows import ( + ProteinDjFormData, + WorkflowFormData, + WorkflowLaunchForm, + WorkflowUserDetails, +) from app.services.proteindj_config import ( get_proteindj_config_profiles, get_proteindj_config_text, @@ -108,11 +113,6 @@ def seqera_env(monkeypatch): # ============================================================================= -def test_get_proteindj_default_params_only_outdir(): - params = get_proteindj_default_params("s3://bucket/out") - assert params == {"outdir": "s3://bucket/out"} - - def test_get_proteindj_default_params_all_fields(): params = get_proteindj_default_params( "s3://bucket/out", @@ -130,9 +130,9 @@ def test_get_proteindj_default_params_all_fields(): } -def test_get_proteindj_default_params_partial_fields(): - params = get_proteindj_default_params("s3://bucket/out", num_designs=3) - assert params == {"outdir": "s3://bucket/out", "num_designs": 3} +def test_get_proteindj_default_params_missing_required_field_raises(): + with pytest.raises(TypeError): + get_proteindj_default_params("s3://bucket/out", num_designs=3) # type: ignore[call-arg] # ============================================================================= @@ -189,20 +189,17 @@ def test_get_proteindj_config_text_contains_base_config(): # ============================================================================= -def test_design_length_both_present(): - assert _design_length(_form_data(min_length=100, max_length=150)) == "100-150" - - -def test_design_length_missing_min(): - assert _design_length(_form_data(max_length=150)) is None - - -def test_design_length_missing_max(): - assert _design_length(_form_data(min_length=100)) is None - - -def test_design_length_missing_both(): - assert _design_length(_form_data()) is None +def test_design_length_formats_min_max_range(): + fields = ProteinDjFormData( + workflow="de-novo-design", + tool="rfdiffusion", + starting_pdb="s3://bucket/in.pdb", + target_hotspot_residues="A20", + number_of_final_designs=5, + min_length=100, + max_length=150, + ) + assert _design_length(fields) == "100-150" # ============================================================================= @@ -302,7 +299,13 @@ async def test_prepare_proteindj_workflow_appends_custom_params_text( pipeline="https://github.com/org/proteindj", config_path="/fake/proteindj.config", output_id="run-output-id", - form_data=_form_data(), + form_data=_form_data( + starting_pdb="s3://my-bucket/inputs/test.pdb", + target_hotspot_residues="A20,A21", + number_of_final_designs=5, + min_length=100, + max_length=150, + ), user_details=_USER_DETAILS, ) @@ -370,6 +373,101 @@ async def test_prepare_proteindj_workflow_empty_output_id(seqera_env): ) +@pytest.mark.anyio +async def test_prepare_proteindj_workflow_missing_starting_pdb(seqera_env): + form = _make_launch_form() + with ( + _mock_proteindj_db_context() as (db_session, workflow_run, *_), + pytest.raises(SeqeraConfigurationError, match="starting_pdb"), + ): + await prepare_proteindj_workflow( + form=form, + db_session=db_session, + workflow_run=workflow_run, + pipeline="https://github.com/org/proteindj", + config_path="/fake/proteindj.config", + output_id="run-output-id", + form_data=_form_data( + target_hotspot_residues="A20,A21", + number_of_final_designs=5, + min_length=100, + max_length=150, + ), + user_details=_USER_DETAILS, + ) + + +@pytest.mark.anyio +async def test_prepare_proteindj_workflow_missing_hotspot_residues(seqera_env): + form = _make_launch_form() + with ( + _mock_proteindj_db_context() as (db_session, workflow_run, *_), + pytest.raises(SeqeraConfigurationError, match="target_hotspot_residues"), + ): + await prepare_proteindj_workflow( + form=form, + db_session=db_session, + workflow_run=workflow_run, + pipeline="https://github.com/org/proteindj", + config_path="/fake/proteindj.config", + output_id="run-output-id", + form_data=_form_data( + starting_pdb="s3://my-bucket/inputs/test.pdb", + number_of_final_designs=5, + min_length=100, + max_length=150, + ), + user_details=_USER_DETAILS, + ) + + +@pytest.mark.anyio +async def test_prepare_proteindj_workflow_missing_num_designs(seqera_env): + form = _make_launch_form() + with ( + _mock_proteindj_db_context() as (db_session, workflow_run, *_), + pytest.raises(SeqeraConfigurationError, match="number_of_final_designs"), + ): + await prepare_proteindj_workflow( + form=form, + db_session=db_session, + workflow_run=workflow_run, + pipeline="https://github.com/org/proteindj", + config_path="/fake/proteindj.config", + output_id="run-output-id", + form_data=_form_data( + starting_pdb="s3://my-bucket/inputs/test.pdb", + target_hotspot_residues="A20,A21", + min_length=100, + max_length=150, + ), + user_details=_USER_DETAILS, + ) + + +@pytest.mark.anyio +async def test_prepare_proteindj_workflow_missing_design_length(seqera_env): + form = _make_launch_form() + with ( + _mock_proteindj_db_context() as (db_session, workflow_run, *_), + pytest.raises(SeqeraConfigurationError, match="min_length"), + ): + await prepare_proteindj_workflow( + form=form, + db_session=db_session, + workflow_run=workflow_run, + pipeline="https://github.com/org/proteindj", + config_path="/fake/proteindj.config", + output_id="run-output-id", + form_data=_form_data( + starting_pdb="s3://my-bucket/inputs/test.pdb", + target_hotspot_residues="A20,A21", + number_of_final_designs=5, + ), + user_details=_USER_DETAILS, + ) + + # ============================================================================= # Tests for launch_proteindj_workflow() # ============================================================================= From 36846cef2923ab2f29a1d10e80dc4890db0d3a30 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Wed, 22 Jul 2026 10:25:04 +1000 Subject: [PATCH 41/63] fix: refactor error handling --- app/services/proteindj_executor.py | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) diff --git a/app/services/proteindj_executor.py b/app/services/proteindj_executor.py index 0a199ee..45309ce 100644 --- a/app/services/proteindj_executor.py +++ b/app/services/proteindj_executor.py @@ -45,10 +45,13 @@ def _parse_proteindj_form_data(form_data: WorkflowFormData) -> ProteinDjFormData try: return ProteinDjFormData.model_validate(form_data.model_dump()) except ValidationError as exc: - missing = next( - (str(e["loc"][-1]) for e in exc.errors() if e.get("loc")), - "formData", - ) + missing = "formData" + for error in exc.errors(): + loc = error.get("loc") + if loc: + *_, field_name = loc # last element of the location path is the field name + missing = str(field_name) + break raise SeqeraConfigurationError( f"'{missing}' is required in formData for ProteinDJ workflow launch" ) from exc From 22d1a12a20c7c52c95e2f9d4d2f2a1d0ca6d014b Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Wed, 22 Jul 2026 13:26:39 +1000 Subject: [PATCH 42/63] feat: additional proteindj form data validation --- app/schemas/workflows.py | 39 +++++++++++++++++--- tests/test_proteindj_coverage.py | 62 ++++++++++++++++++++++++++++++++ 2 files changed, 96 insertions(+), 5 deletions(-) diff --git a/app/schemas/workflows.py b/app/schemas/workflows.py index 3b76b0b..f0a8be2 100644 --- a/app/schemas/workflows.py +++ b/app/schemas/workflows.py @@ -6,7 +6,7 @@ from enum import StrEnum from typing import Any, Literal -from pydantic import BaseModel, ConfigDict, Field, field_validator +from pydantic import BaseModel, ConfigDict, Field, field_validator, model_validator WorkflowName = Literal[ "single-prediction", "de-novo-design", "bulk-prediction", "interaction-screening" @@ -111,6 +111,11 @@ class WispsFormData(WorkflowFormData): ) +MAX_HOTSPOT_RESIDUES = 8 +MIN_DESIGN_LENGTH = 65 +MAX_DESIGN_LENGTH = 150 + + class ProteinDjFormData(WorkflowFormData): """Form data for the ProteinDJ (rfdiffusion) de-novo-design workflow.""" @@ -119,17 +124,41 @@ class ProteinDjFormData(WorkflowFormData): ..., description="Comma-separated hotspot residues, e.g. 'A20,A21'" ) number_of_final_designs: int = Field(..., ge=1, description="Number of designs to generate") - min_length: int = Field(..., description="Minimum binder length") - max_length: int = Field(..., description="Maximum binder length") + min_length: int = Field( + ..., ge=MIN_DESIGN_LENGTH, description="Minimum binder length" + ) + max_length: int = Field( + ..., le=MAX_DESIGN_LENGTH, description="Maximum binder length" + ) - @field_validator("starting_pdb", "target_hotspot_residues") + @field_validator("starting_pdb") @classmethod - def _not_blank(cls, value: str) -> str: + def _validate_starting_pdb(cls, value: str) -> str: stripped = value.strip() if not stripped: raise ValueError("This field is required") return stripped + @field_validator("target_hotspot_residues") + @classmethod + def _validate_hotspot_residues(cls, value: str) -> str: + stripped = value.strip() + if not stripped: + raise ValueError("This field is required") + residue_count = len([token for token in stripped.split(",") if token.strip()]) + if residue_count > MAX_HOTSPOT_RESIDUES: + raise ValueError( + f"Too many hotspot residues ({residue_count}); only up to " + f"{MAX_HOTSPOT_RESIDUES} are supported" + ) + return stripped + + @model_validator(mode="after") + def _validate_length_range(self) -> "ProteinDjFormData": + if self.min_length > self.max_length: + raise ValueError("min_length must not exceed max_length") + return self + class WorkflowLaunchPayload(BaseModel): model_config = ConfigDict(extra="forbid") diff --git a/tests/test_proteindj_coverage.py b/tests/test_proteindj_coverage.py index 6204a47..9d3b69d 100644 --- a/tests/test_proteindj_coverage.py +++ b/tests/test_proteindj_coverage.py @@ -108,6 +108,68 @@ def seqera_env(monkeypatch): monkeypatch.setenv("AWS_SECRET_ACCESS_KEY", "test_secret") +def _proteindj_form_kwargs(**overrides) -> dict: + defaults = { + "workflow": "de-novo-design", + "tool": "rfdiffusion", + "starting_pdb": "s3://bucket/in.pdb", + "target_hotspot_residues": "A20,A21", + "number_of_final_designs": 5, + "min_length": 100, + "max_length": 150, + } + defaults.update(overrides) + return defaults + + +# ============================================================================= +# Tests for ProteinDjFormData validation +# ============================================================================= + + +def test_proteindj_form_data_accepts_valid_fields(): + fields = ProteinDjFormData(**_proteindj_form_kwargs()) + assert fields.min_length == 100 + assert fields.max_length == 150 + + +def test_proteindj_form_data_rejects_too_many_hotspot_residues(): + with pytest.raises(ValueError, match="Too many hotspot residues"): + ProteinDjFormData( + **_proteindj_form_kwargs( + target_hotspot_residues="A1,A2,A3,A4,A5,A6,A7,A8,A9" + ) + ) + + +def test_proteindj_form_data_allows_exactly_max_hotspot_residues(): + fields = ProteinDjFormData( + **_proteindj_form_kwargs(target_hotspot_residues="A1,A2,A3,A4,A5,A6,A7,A8") + ) + assert fields.target_hotspot_residues == "A1,A2,A3,A4,A5,A6,A7,A8" + + +def test_proteindj_form_data_rejects_min_length_below_floor(): + with pytest.raises(ValueError, match="min_length"): + ProteinDjFormData(**_proteindj_form_kwargs(min_length=64, max_length=150)) + + +def test_proteindj_form_data_rejects_max_length_above_ceiling(): + with pytest.raises(ValueError, match="max_length"): + ProteinDjFormData(**_proteindj_form_kwargs(min_length=65, max_length=151)) + + +def test_proteindj_form_data_rejects_min_length_above_max_length(): + with pytest.raises(ValueError, match="min_length must not exceed max_length"): + ProteinDjFormData(**_proteindj_form_kwargs(min_length=150, max_length=100)) + + +def test_proteindj_form_data_allows_boundary_lengths(): + fields = ProteinDjFormData(**_proteindj_form_kwargs(min_length=65, max_length=150)) + assert fields.min_length == 65 + assert fields.max_length == 150 + + # ============================================================================= # Tests for get_proteindj_default_params() # ============================================================================= From cd3d8eb1746d164f66942ce2e2e879ad73a7564d Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Wed, 22 Jul 2026 13:28:28 +1000 Subject: [PATCH 43/63] chore: lint --- app/schemas/workflows.py | 10 +++------- tests/test_proteindj_coverage.py | 4 +--- 2 files changed, 4 insertions(+), 10 deletions(-) diff --git a/app/schemas/workflows.py b/app/schemas/workflows.py index f0a8be2..7ebae53 100644 --- a/app/schemas/workflows.py +++ b/app/schemas/workflows.py @@ -124,12 +124,8 @@ class ProteinDjFormData(WorkflowFormData): ..., description="Comma-separated hotspot residues, e.g. 'A20,A21'" ) number_of_final_designs: int = Field(..., ge=1, description="Number of designs to generate") - min_length: int = Field( - ..., ge=MIN_DESIGN_LENGTH, description="Minimum binder length" - ) - max_length: int = Field( - ..., le=MAX_DESIGN_LENGTH, description="Maximum binder length" - ) + min_length: int = Field(..., ge=MIN_DESIGN_LENGTH, description="Minimum binder length") + max_length: int = Field(..., le=MAX_DESIGN_LENGTH, description="Maximum binder length") @field_validator("starting_pdb") @classmethod @@ -154,7 +150,7 @@ def _validate_hotspot_residues(cls, value: str) -> str: return stripped @model_validator(mode="after") - def _validate_length_range(self) -> "ProteinDjFormData": + def _validate_length_range(self) -> ProteinDjFormData: if self.min_length > self.max_length: raise ValueError("min_length must not exceed max_length") return self diff --git a/tests/test_proteindj_coverage.py b/tests/test_proteindj_coverage.py index 9d3b69d..2c9efa5 100644 --- a/tests/test_proteindj_coverage.py +++ b/tests/test_proteindj_coverage.py @@ -136,9 +136,7 @@ def test_proteindj_form_data_accepts_valid_fields(): def test_proteindj_form_data_rejects_too_many_hotspot_residues(): with pytest.raises(ValueError, match="Too many hotspot residues"): ProteinDjFormData( - **_proteindj_form_kwargs( - target_hotspot_residues="A1,A2,A3,A4,A5,A6,A7,A8,A9" - ) + **_proteindj_form_kwargs(target_hotspot_residues="A1,A2,A3,A4,A5,A6,A7,A8,A9") ) From 18b26996b3ac0122e7b6a34efea558da7426b39e Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Wed, 22 Jul 2026 13:54:35 +1000 Subject: [PATCH 44/63] feat: encode email and ip for cluster options --- app/services/bindflow_config.py | 6 +++--- app/services/cluster_utils.py | 6 +++--- app/services/proteindj_config.py | 6 +++--- app/services/proteinfold_config.py | 6 +++--- app/services/wisps_config.py | 6 +++--- 5 files changed, 15 insertions(+), 15 deletions(-) diff --git a/app/services/bindflow_config.py b/app/services/bindflow_config.py index 359342f..68dec86 100644 --- a/app/services/bindflow_config.py +++ b/app/services/bindflow_config.py @@ -5,7 +5,7 @@ from typing import Any from ..schemas.workflows import WorkflowUserDetails -from .cluster_utils import GADI_PROJECT, encode_ip +from .cluster_utils import GADI_PROJECT, encode_value from .workflow_config_fetcher import fetch_workflow_config @@ -32,9 +32,9 @@ def get_bindflow_config_text( base = fetch_workflow_config(config_file_path) account = ( - f"{user_details.user_email}:{encode_ip(user_details.ip_address)}" + f"{encode_value(user_details.user_email)}:{encode_value(user_details.ip_address)}" if user_details.ip_address - else user_details.user_email + else encode_value(user_details.user_email) ) cluster_opts = f"-A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' diff --git a/app/services/cluster_utils.py b/app/services/cluster_utils.py index 6d0af78..ae15a48 100644 --- a/app/services/cluster_utils.py +++ b/app/services/cluster_utils.py @@ -8,6 +8,6 @@ GADI_PROJECT: str = os.getenv("GADI_PROJECT", "yz52") -def encode_ip(ip_address: str) -> str: - """Return the base64 encoding of an IP address string.""" - return base64.b64encode(ip_address.encode()).decode() +def encode_value(value: str) -> str: + """Return the base64 encoding of a string.""" + return base64.b64encode(value.encode()).decode() diff --git a/app/services/proteindj_config.py b/app/services/proteindj_config.py index bc83991..37718d1 100644 --- a/app/services/proteindj_config.py +++ b/app/services/proteindj_config.py @@ -5,7 +5,7 @@ from typing import Any from ..schemas.workflows import WorkflowUserDetails -from .cluster_utils import encode_ip +from .cluster_utils import encode_value from .workflow_config_fetcher import fetch_workflow_config @@ -44,9 +44,9 @@ def get_proteindj_config_text( base = fetch_workflow_config(config_file_path) account = ( - f"{user_details.user_email}:{encode_ip(user_details.ip_address)}" + f"{encode_value(user_details.user_email)}:{encode_value(user_details.ip_address)}" if user_details.ip_address - else user_details.user_email + else encode_value(user_details.user_email) ) cluster_opts = f"-A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' diff --git a/app/services/proteinfold_config.py b/app/services/proteinfold_config.py index 4241c49..aad514d 100644 --- a/app/services/proteinfold_config.py +++ b/app/services/proteinfold_config.py @@ -5,7 +5,7 @@ from typing import Any from ..schemas.workflows import WorkflowUserDetails -from .cluster_utils import GADI_PROJECT, encode_ip +from .cluster_utils import GADI_PROJECT, encode_value from .workflow_config_fetcher import fetch_workflow_config @@ -30,9 +30,9 @@ def get_proteinfold_config_text( base = fetch_workflow_config(config_file_path) account = ( - f"{user_details.user_email}:{encode_ip(user_details.ip_address)}" + f"{encode_value(user_details.user_email)}:{encode_value(user_details.ip_address)}" if user_details.ip_address - else user_details.user_email + else encode_value(user_details.user_email) ) cluster_opts = f"-P {GADI_PROJECT} -A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' diff --git a/app/services/wisps_config.py b/app/services/wisps_config.py index 6f18c65..94774bd 100644 --- a/app/services/wisps_config.py +++ b/app/services/wisps_config.py @@ -5,7 +5,7 @@ from typing import Any, Literal from ..schemas.workflows import WorkflowUserDetails -from .cluster_utils import GADI_PROJECT, encode_ip +from .cluster_utils import GADI_PROJECT, encode_value from .workflow_config_fetcher import fetch_workflow_config WispsMode = Literal["g1-g2", "manual"] @@ -50,9 +50,9 @@ def get_wisps_config_text( base = fetch_workflow_config(config_file_path) account = ( - f"{user_details.user_email}:{encode_ip(user_details.ip_address)}" + f"{encode_value(user_details.user_email)}:{encode_value(user_details.ip_address)}" if user_details.ip_address - else user_details.user_email + else encode_value(user_details.user_email) ) cluster_opts = f"-P {GADI_PROJECT} -A {account}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' From 766acff8f87b77e2b327b1ddf44843bb2e49821f Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Wed, 22 Jul 2026 13:54:57 +1000 Subject: [PATCH 45/63] test: fix encoded email in cluster options --- tests/test_proteindj_coverage.py | 8 ++++---- tests/test_proteinfold_coverage.py | 8 ++++---- tests/test_services_bindflow_config.py | 6 +++--- tests/test_services_cluster_utils.py | 19 ++++++++++++------- tests/test_wisps_coverage.py | 8 ++++---- 5 files changed, 27 insertions(+), 22 deletions(-) diff --git a/tests/test_proteindj_coverage.py b/tests/test_proteindj_coverage.py index 2c9efa5..1c863b1 100644 --- a/tests/test_proteindj_coverage.py +++ b/tests/test_proteindj_coverage.py @@ -221,20 +221,20 @@ def test_get_proteindj_config_text_appends_process_block(): assert "clusterOptions" in result -def test_get_proteindj_config_text_contains_email_and_encoded_ip(): +def test_get_proteindj_config_text_contains_encoded_email_and_encoded_ip(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_proteindj_config_text("/fake/proteindj.config", user_details=_USER_DETAILS) - assert "user@ex.com" in result + assert "dXNlckBleC5jb20=" in result assert "MS4yLjMuNA==" in result -def test_get_proteindj_config_text_without_ip_uses_email_only(): +def test_get_proteindj_config_text_without_ip_uses_encoded_email_only(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_proteindj_config_text( "/fake/proteindj.config", user_details=_USER_DETAILS.model_copy(update={"ip_address": ""}), ) - assert "-A user@ex.com" in result + assert "-A dXNlckBleC5jb20=" in result assert ":" not in result.split("clusterOptions = ")[1] diff --git a/tests/test_proteinfold_coverage.py b/tests/test_proteinfold_coverage.py index b96a77b..e8918a2 100644 --- a/tests/test_proteinfold_coverage.py +++ b/tests/test_proteinfold_coverage.py @@ -547,23 +547,23 @@ def test_get_proteinfold_config_text_appends_process_block(): assert "clusterOptions" in result -def test_get_proteinfold_config_text_contains_email_and_encoded_ip(): +def test_get_proteinfold_config_text_contains_encoded_email_and_encoded_ip(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_proteinfold_config_text( "/fake/proteinfold.config", user_details=_USER_DETAILS, ) - assert "user@ex.com" in result + assert "dXNlckBleC5jb20=" in result assert "MS4yLjMuNA==" in result -def test_get_proteinfold_config_text_without_ip_uses_email_only(): +def test_get_proteinfold_config_text_without_ip_uses_encoded_email_only(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_proteinfold_config_text( "/fake/proteinfold.config", user_details=_USER_DETAILS.model_copy(update={"ip_address": ""}), ) - assert "-A user@ex.com" in result + assert "-A dXNlckBleC5jb20=" in result assert ":" not in result.split("clusterOptions = ")[1] diff --git a/tests/test_services_bindflow_config.py b/tests/test_services_bindflow_config.py index 95935c0..f1f69d3 100644 --- a/tests/test_services_bindflow_config.py +++ b/tests/test_services_bindflow_config.py @@ -145,7 +145,7 @@ def test_get_bindflow_config_text_interpolates_email(): "/fake/bindflow.config", user_details=_user_details("alice@example.com"), ) - assert "-A alice@example.com" in result + assert "-A YWxpY2VAZXhhbXBsZS5jb20=" in result def test_get_bindflow_config_text_without_ip_address_omits_encoding(): @@ -154,7 +154,7 @@ def test_get_bindflow_config_text_without_ip_address_omits_encoding(): "/fake/bindflow.config", user_details=_user_details("user@example.com"), ) - assert "-A user@example.com" in result + assert "-A dXNlckBleGFtcGxlLmNvbQ==" in result assert ":" not in result.split("clusterOptions = ")[1] @@ -164,7 +164,7 @@ def test_get_bindflow_config_text_with_ip_address_appends_encoded_ip(): "/fake/bindflow.config", user_details=_user_details("user@example.com", ip_address="1.2.3.4"), ) - assert "-A user@example.com:MS4yLjMuNA==" in result + assert "-A dXNlckBleGFtcGxlLmNvbQ==:MS4yLjMuNA==" in result def test_get_bindflow_config_text_url_fetching(): diff --git a/tests/test_services_cluster_utils.py b/tests/test_services_cluster_utils.py index fad9453..92fe94a 100644 --- a/tests/test_services_cluster_utils.py +++ b/tests/test_services_cluster_utils.py @@ -7,20 +7,25 @@ import importlib from app.services import cluster_utils -from app.services.cluster_utils import encode_ip +from app.services.cluster_utils import encode_value -def test_encode_ip_returns_base64_of_ip_string(): - assert encode_ip("1.2.3.4") == base64.b64encode(b"1.2.3.4").decode() +def test_encode_value_returns_base64_of_ip_string(): + assert encode_value("1.2.3.4") == base64.b64encode(b"1.2.3.4").decode() -def test_encode_ip_empty_string(): - assert encode_ip("") == "" +def test_encode_value_empty_string(): + assert encode_value("") == "" -def test_encode_ip_ipv6_address(): +def test_encode_value_ipv6_address(): ip = "2001:db8::1" - assert encode_ip(ip) == base64.b64encode(ip.encode()).decode() + assert encode_value(ip) == base64.b64encode(ip.encode()).decode() + + +def test_encode_value_email_string(): + email = "user@example.com" + assert encode_value(email) == base64.b64encode(email.encode()).decode() def test_gadi_project_defaults_to_yz52(monkeypatch): diff --git a/tests/test_wisps_coverage.py b/tests/test_wisps_coverage.py index f842c61..dedf83a 100644 --- a/tests/test_wisps_coverage.py +++ b/tests/test_wisps_coverage.py @@ -262,23 +262,23 @@ def test_get_wisps_config_text_appends_process_block(): assert "clusterOptions" in result -def test_get_wisps_config_text_contains_email_and_encoded_ip(): +def test_get_wisps_config_text_contains_encoded_email_and_encoded_ip(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_wisps_config_text( config_file_path="/fake/path.config", user_details=_USER_DETAILS, ) - assert "user@ex.com" in result + assert "dXNlckBleC5jb20=" in result assert "MS4yLjMuNA==" in result -def test_get_wisps_config_text_without_ip_uses_email_only(): +def test_get_wisps_config_text_without_ip_uses_encoded_email_only(): with patch("builtins.open", mock_open(read_data="base_config")): result = get_wisps_config_text( config_file_path="/fake/path.config", user_details=_USER_DETAILS.model_copy(update={"ip_address": ""}), ) - assert "-A user@ex.com" in result + assert "-A dXNlckBleC5jb20=" in result assert ":" not in result.split("clusterOptions = ")[1] From b26d03204b5de5e7d73d721f434f2fd3562f6283 Mon Sep 17 00:00:00 2001 From: Minh Vu Date: Wed, 22 Jul 2026 14:19:32 +1000 Subject: [PATCH 46/63] refactor: Pydantic types --- app/schemas/workflows.py | 24 ++++++++++++++++-------- tests/test_schemas.py | 17 +++++++++++++---- 2 files changed, 29 insertions(+), 12 deletions(-) diff --git a/app/schemas/workflows.py b/app/schemas/workflows.py index a549761..490ff22 100644 --- a/app/schemas/workflows.py +++ b/app/schemas/workflows.py @@ -5,14 +5,24 @@ import re from datetime import datetime from enum import StrEnum -from typing import Any, Literal +from typing import Annotated, Any, Literal -from pydantic import BaseModel, ConfigDict, Field, field_validator +from pydantic import ( + BaseModel, + ConfigDict, + Field, + PositiveInt, + StringConstraints, + field_validator, +) WorkflowName = Literal[ "single-prediction", "de-novo-design", "bulk-prediction", "interaction-screening" ] WorkflowTool = Literal["alphafold2", "bindcraft", "boltz", "colabfold", "rfdiffusion"] +MoleculeType = Literal["protein", "rna", "dna", "ligand", "ccd"] + +NonEmptyStr = Annotated[str, StringConstraints(strip_whitespace=True, min_length=1)] SINGLE_PREDICTION_MAX_ENTITIES = 52 SINGLE_PREDICTION_LIGAND_SIZE = 30 @@ -126,10 +136,10 @@ class SinglePredictionEntity(BaseModel): model_config = ConfigDict(extra="allow") - id: str | None = None - moleculeType: str - copyNumber: int = 1 - sequence: str = "" + id: NonEmptyStr | None = None + moleculeType: MoleculeType + copyNumber: PositiveInt = 1 + sequence: Annotated[str, StringConstraints(strip_whitespace=True)] = "" def _entity_prediction_size(entity: SinglePredictionEntity) -> int: @@ -163,8 +173,6 @@ def validate_single_prediction_entities( total_size = 0 has_protein = False for entity in entities: - if entity.copyNumber < 1: - raise ValueError("Each entity must have a copy number of at least 1.") total_copies += entity.copyNumber if entity.moleculeType == "protein": has_protein = True diff --git a/tests/test_schemas.py b/tests/test_schemas.py index 895e9ad..3cff85f 100644 --- a/tests/test_schemas.py +++ b/tests/test_schemas.py @@ -507,10 +507,19 @@ def test_validate_single_prediction_requires_protein(): validate_single_prediction_entities([dna], "boltz") -def test_validate_single_prediction_rejects_copy_number_below_one(): - entity = SinglePredictionEntity(id="p", moleculeType="protein", copyNumber=0, sequence="ACD") - with pytest.raises(ValueError, match="copy number of at least 1"): - validate_single_prediction_entities([entity], "colabfold") +def test_single_prediction_entity_rejects_copy_number_below_one(): + with pytest.raises(ValidationError): + SinglePredictionEntity(id="p", moleculeType="protein", copyNumber=0, sequence="ACD") + + +def test_single_prediction_entity_rejects_unknown_molecule_type(): + with pytest.raises(ValidationError): + SinglePredictionEntity(id="p", moleculeType="peptide", copyNumber=1, sequence="ACD") + + +def test_single_prediction_entity_rejects_empty_id(): + with pytest.raises(ValidationError): + SinglePredictionEntity(id="", moleculeType="protein", copyNumber=1, sequence="ACD") def test_validate_single_prediction_ligand_uses_fixed_size(): From 8d90f2a52d1a86232ab27d0550244b6ab75f6a32 Mon Sep 17 00:00:00 2001 From: Minh Vu Date: Wed, 22 Jul 2026 15:35:20 +1000 Subject: [PATCH 47/63] refactor: rename 'alphafold2_random_seed' to 'random_seed' in tool parameters --- app/services/proteinfold_executor.py | 2 +- tests/test_proteinfold_coverage.py | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/app/services/proteinfold_executor.py b/app/services/proteinfold_executor.py index 59002fb..bf73a50 100644 --- a/app/services/proteinfold_executor.py +++ b/app/services/proteinfold_executor.py @@ -30,7 +30,7 @@ # Params forwarded from the frontend's Tool Settings (step 2) _TOOL_PARAM_KEYS = frozenset( { - "alphafold2_random_seed", + "random_seed", "alphafold2_full_dbs", "colabfold_num_recycles", "colabfold_use_templates", diff --git a/tests/test_proteinfold_coverage.py b/tests/test_proteinfold_coverage.py index b96a77b..e8c728b 100644 --- a/tests/test_proteinfold_coverage.py +++ b/tests/test_proteinfold_coverage.py @@ -141,8 +141,8 @@ def test_tool_params_with_int(): def test_tool_params_with_str(): - result = _tool_params(_form_data(alphafold2_random_seed="42")) - assert result == {"alphafold2_random_seed": "42"} + result = _tool_params(_form_data(random_seed="42")) + assert result == {"random_seed": "42"} def test_tool_params_none_value_excluded(): From 1a9806e30405e6e29274a0ae3f0e6af341ebf3a7 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Wed, 22 Jul 2026 15:45:40 +1000 Subject: [PATCH 48/63] refactor: encode user email and IP address --- app/schemas/workflows.py | 13 +++++++++++++ app/services/bindflow_config.py | 9 ++------- app/services/cluster_utils.py | 6 ------ app/services/proteindj_config.py | 8 +------- app/services/proteinfold_config.py | 9 ++------- app/services/wisps_config.py | 9 ++------- tests/test_schemas.py | 11 +++++++++++ tests/test_services_cluster_utils.py | 20 -------------------- 8 files changed, 31 insertions(+), 54 deletions(-) diff --git a/app/schemas/workflows.py b/app/schemas/workflows.py index 7ebae53..8831d49 100644 --- a/app/schemas/workflows.py +++ b/app/schemas/workflows.py @@ -2,6 +2,7 @@ from __future__ import annotations +import base64 from datetime import datetime from enum import StrEnum from typing import Any, Literal @@ -99,6 +100,18 @@ class WorkflowUserDetails(BaseModel): user_email: str = Field(..., description="Email address of the user") ip_address: str = Field(..., description="IP address of the user") + def get_encoded_account_details(self) -> str: + """Return the `-A` account string for cluster job submission. + + Encodes email and IP as base64 (colon-joined) so the account string + never leaks the user's raw email/IP into cluster logs or job metadata. + """ + encoded_email = base64.b64encode(self.user_email.encode()).decode() + if not self.ip_address: + return encoded_email + encoded_ip = base64.b64encode(self.ip_address.encode()).decode() + return f"{encoded_email}:{encoded_ip}" + class WispsFormData(WorkflowFormData): """Form data for WISPS workflows (interaction-screening, bulk-prediction).""" diff --git a/app/services/bindflow_config.py b/app/services/bindflow_config.py index 68dec86..8f831d5 100644 --- a/app/services/bindflow_config.py +++ b/app/services/bindflow_config.py @@ -5,7 +5,7 @@ from typing import Any from ..schemas.workflows import WorkflowUserDetails -from .cluster_utils import GADI_PROJECT, encode_value +from .cluster_utils import GADI_PROJECT from .workflow_config_fetcher import fetch_workflow_config @@ -31,11 +31,6 @@ def get_bindflow_config_text( """Read bindflow base config and append a process override block with runtime values.""" base = fetch_workflow_config(config_file_path) - account = ( - f"{encode_value(user_details.user_email)}:{encode_value(user_details.ip_address)}" - if user_details.ip_address - else encode_value(user_details.user_email) - ) - cluster_opts = f"-A {account}" + cluster_opts = f"-A {user_details.get_encoded_account_details()}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override diff --git a/app/services/cluster_utils.py b/app/services/cluster_utils.py index ae15a48..b7d3da2 100644 --- a/app/services/cluster_utils.py +++ b/app/services/cluster_utils.py @@ -2,12 +2,6 @@ from __future__ import annotations -import base64 import os GADI_PROJECT: str = os.getenv("GADI_PROJECT", "yz52") - - -def encode_value(value: str) -> str: - """Return the base64 encoding of a string.""" - return base64.b64encode(value.encode()).decode() diff --git a/app/services/proteindj_config.py b/app/services/proteindj_config.py index 37718d1..1f73557 100644 --- a/app/services/proteindj_config.py +++ b/app/services/proteindj_config.py @@ -5,7 +5,6 @@ from typing import Any from ..schemas.workflows import WorkflowUserDetails -from .cluster_utils import encode_value from .workflow_config_fetcher import fetch_workflow_config @@ -43,11 +42,6 @@ def get_proteindj_config_text( """Read proteindj base config and append a process override block with runtime values.""" base = fetch_workflow_config(config_file_path) - account = ( - f"{encode_value(user_details.user_email)}:{encode_value(user_details.ip_address)}" - if user_details.ip_address - else encode_value(user_details.user_email) - ) - cluster_opts = f"-A {account}" + cluster_opts = f"-A {user_details.get_encoded_account_details()}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override diff --git a/app/services/proteinfold_config.py b/app/services/proteinfold_config.py index aad514d..b836a2a 100644 --- a/app/services/proteinfold_config.py +++ b/app/services/proteinfold_config.py @@ -5,7 +5,7 @@ from typing import Any from ..schemas.workflows import WorkflowUserDetails -from .cluster_utils import GADI_PROJECT, encode_value +from .cluster_utils import GADI_PROJECT from .workflow_config_fetcher import fetch_workflow_config @@ -29,11 +29,6 @@ def get_proteinfold_config_text( """Read proteinfold base config and append a process override block with runtime values.""" base = fetch_workflow_config(config_file_path) - account = ( - f"{encode_value(user_details.user_email)}:{encode_value(user_details.ip_address)}" - if user_details.ip_address - else encode_value(user_details.user_email) - ) - cluster_opts = f"-P {GADI_PROJECT} -A {account}" + cluster_opts = f"-P {GADI_PROJECT} -A {user_details.get_encoded_account_details()}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override diff --git a/app/services/wisps_config.py b/app/services/wisps_config.py index 94774bd..d265df5 100644 --- a/app/services/wisps_config.py +++ b/app/services/wisps_config.py @@ -5,7 +5,7 @@ from typing import Any, Literal from ..schemas.workflows import WorkflowUserDetails -from .cluster_utils import GADI_PROJECT, encode_value +from .cluster_utils import GADI_PROJECT from .workflow_config_fetcher import fetch_workflow_config WispsMode = Literal["g1-g2", "manual"] @@ -49,11 +49,6 @@ def get_wisps_config_text( """ base = fetch_workflow_config(config_file_path) - account = ( - f"{encode_value(user_details.user_email)}:{encode_value(user_details.ip_address)}" - if user_details.ip_address - else encode_value(user_details.user_email) - ) - cluster_opts = f"-P {GADI_PROJECT} -A {account}" + cluster_opts = f"-P {GADI_PROJECT} -A {user_details.get_encoded_account_details()}" override = f'\nprocess {{\n clusterOptions = "{cluster_opts}"\n}}\n' return base + override diff --git a/tests/test_schemas.py b/tests/test_schemas.py index fbe5200..461a101 100644 --- a/tests/test_schemas.py +++ b/tests/test_schemas.py @@ -23,6 +23,7 @@ WorkflowLaunchForm, WorkflowLaunchPayload, WorkflowLaunchResponse, + WorkflowUserDetails, map_pipeline_status_to_ui, ) @@ -462,3 +463,13 @@ def test_interaction_screening_request_extra_fields_forbidden(): runId="run-1", extra="bad", ) + + +def test_get_encoded_account_details_encodes_email_and_ip(): + details = WorkflowUserDetails(user_email="user@example.com", ip_address="1.2.3.4") + assert details.get_encoded_account_details() == "dXNlckBleGFtcGxlLmNvbQ==:MS4yLjMuNA==" + + +def test_get_encoded_account_details_without_ip_encodes_email_only(): + details = WorkflowUserDetails(user_email="user@example.com", ip_address="") + assert details.get_encoded_account_details() == "dXNlckBleGFtcGxlLmNvbQ==" diff --git a/tests/test_services_cluster_utils.py b/tests/test_services_cluster_utils.py index 92fe94a..f4fc80c 100644 --- a/tests/test_services_cluster_utils.py +++ b/tests/test_services_cluster_utils.py @@ -3,29 +3,9 @@ # pylint: disable=missing-function-docstring from __future__ import annotations -import base64 import importlib from app.services import cluster_utils -from app.services.cluster_utils import encode_value - - -def test_encode_value_returns_base64_of_ip_string(): - assert encode_value("1.2.3.4") == base64.b64encode(b"1.2.3.4").decode() - - -def test_encode_value_empty_string(): - assert encode_value("") == "" - - -def test_encode_value_ipv6_address(): - ip = "2001:db8::1" - assert encode_value(ip) == base64.b64encode(ip.encode()).decode() - - -def test_encode_value_email_string(): - email = "user@example.com" - assert encode_value(email) == base64.b64encode(email.encode()).decode() def test_gadi_project_defaults_to_yz52(monkeypatch): From adddbe6142428bc2c365016a28aa6d58c3a77e02 Mon Sep 17 00:00:00 2001 From: laclac102 <53896516+vtnphan@users.noreply.github.com> Date: Thu, 23 Jul 2026 09:06:10 +1000 Subject: [PATCH 49/63] chore: lint --- tests/test_schemas.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tests/test_schemas.py b/tests/test_schemas.py index 8403602..a7a5a80 100644 --- a/tests/test_schemas.py +++ b/tests/test_schemas.py @@ -476,6 +476,8 @@ def test_get_encoded_account_details_encodes_email_and_ip(): def test_get_encoded_account_details_without_ip_encodes_email_only(): details = WorkflowUserDetails(user_email="user@example.com", ip_address="") assert details.get_encoded_account_details() == "dXNlckBleGFtcGxlLmNvbQ==" + + # ============================================================================= # Single-prediction entity validation # ============================================================================= From c2b59c9c6eb9f97c02ecf2c5fd6610f083c22698 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 24 Jul 2026 10:14:28 +1000 Subject: [PATCH 50/63] feat: record system status incidents in the db --- ...00_system_status_incidents_a6ee1af2c156.py | 33 ++++ app/db/models/system_status.py | 25 ++- app/routes/system_status.py | 30 +++- app/schemas/health.py | 38 +++++ app/services/health.py | 145 +++++++++++++++++- 5 files changed, 265 insertions(+), 6 deletions(-) create mode 100644 alembic/versions/20260724_094600_system_status_incidents_a6ee1af2c156.py diff --git a/alembic/versions/20260724_094600_system_status_incidents_a6ee1af2c156.py b/alembic/versions/20260724_094600_system_status_incidents_a6ee1af2c156.py new file mode 100644 index 0000000..2d559ce --- /dev/null +++ b/alembic/versions/20260724_094600_system_status_incidents_a6ee1af2c156.py @@ -0,0 +1,33 @@ +"""system_status_incidents""" + +from alembic import op +import sqlalchemy as sa + + +# revision identifiers, used by Alembic. +revision = 'a6ee1af2c156' +down_revision = '14cff803a3b9' +branch_labels = None +depends_on = None + + +def upgrade() -> None: + # ### commands auto generated by Alembic - please adjust! ### + op.create_table('system_status_incidents', + sa.Column('id', sa.Integer(), autoincrement=True, nullable=False), + sa.Column('component', sa.Text(), nullable=False), + sa.Column('status', sa.Text(), nullable=False), + sa.Column('started_at', sa.DateTime(timezone=True), nullable=False), + sa.Column('ended_at', sa.DateTime(timezone=True), nullable=True), + sa.Column('message', sa.Text(), nullable=True), + sa.PrimaryKeyConstraint('id', name=op.f('pk_system_status_incidents')) + ) + op.create_index('ix_system_status_incidents_component_started_at', 'system_status_incidents', ['component', 'started_at'], unique=False) + # ### end Alembic commands ### + + +def downgrade() -> None: + # ### commands auto generated by Alembic - please adjust! ### + op.drop_index('ix_system_status_incidents_component_started_at', table_name='system_status_incidents') + op.drop_table('system_status_incidents') + # ### end Alembic commands ### diff --git a/app/db/models/system_status.py b/app/db/models/system_status.py index b01b9bd..be3aff2 100644 --- a/app/db/models/system_status.py +++ b/app/db/models/system_status.py @@ -1,9 +1,9 @@ -"""Database-backed cache for runtime system health.""" +"""Database-backed cache and history for runtime system health.""" from datetime import UTC, datetime from typing import Any -from sqlalchemy import JSON, DateTime, Text +from sqlalchemy import JSON, DateTime, Index, Integer, Text from sqlalchemy.orm import Mapped, mapped_column from ...schemas.health import SystemStatus @@ -38,3 +38,24 @@ def get_status(self) -> SystemStatus: Return a SystemStatus object from the cache payload. """ return SystemStatus.model_validate(self.payload) + + +class SystemStatusIncident(Base): + """ + One row per downtime period for a monitored component: opened when a probe + first reports ``degraded``/``unhealthy``, closed when it next reports + ``healthy``. Healthy checks are never recorded, so storage scales with the + number of actual outages rather than with polling frequency. + """ + + __tablename__ = "system_status_incidents" + __table_args__ = ( + Index("ix_system_status_incidents_component_started_at", "component", "started_at"), + ) + + id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True) + component: Mapped[str] = mapped_column(Text, nullable=False) + status: Mapped[str] = mapped_column(Text, nullable=False) + started_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False) + ended_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True) + message: Mapped[str | None] = mapped_column(Text, nullable=True) diff --git a/app/routes/system_status.py b/app/routes/system_status.py index 2fd0fcc..1fe478f 100644 --- a/app/routes/system_status.py +++ b/app/routes/system_status.py @@ -12,11 +12,13 @@ from __future__ import annotations +from datetime import UTC, datetime, timedelta + from fastapi import APIRouter, Depends, Query from sqlalchemy.orm import Session from ..db.admin import require_admin_access -from ..schemas.health import SystemStatusAdminResponse +from ..schemas.health import SystemStatusAdminResponse, SystemStatusDowntimeResponse from ..services import health from .dependencies import get_db @@ -25,6 +27,9 @@ dependencies=[Depends(require_admin_access)], ) +# Bounds how far back /system-status/history can be queried. +_MAX_HISTORY_HOURS = 24 * 30 + @router.get("/system-status", response_model=SystemStatusAdminResponse) async def get_admin_system_status( @@ -37,3 +42,26 @@ async def get_admin_system_status( """Return verbose, admin-only runtime health of the submission components.""" status_obj = await health.get_system_status(db, force_refresh=refresh) return SystemStatusAdminResponse.model_validate(health.to_admin_dict(status_obj)) + + +@router.get("/system-status/history", response_model=SystemStatusDowntimeResponse) +async def get_admin_system_status_history( + hours: int = Query( + default=24, + ge=1, + le=_MAX_HISTORY_HOURS, + description="How many hours of downtime history to return", + ), + component: str | None = Query( + default=None, description="Restrict to a single component name" + ), + db: Session = Depends(get_db), +) -> SystemStatusDowntimeResponse: + """Return per-component downtime incidents, so admins can see how frequently + and for how long a component has been degraded/unhealthy over time.""" + until = datetime.now(UTC) + since = until - timedelta(hours=hours) + incidents = health.get_incidents(db, since=since, until=until, component=component) + return SystemStatusDowntimeResponse.model_validate( + health.to_downtime_dict(incidents, since=since, until=until, hours=hours) + ) diff --git a/app/schemas/health.py b/app/schemas/health.py index e95c67f..ee47da3 100644 --- a/app/schemas/health.py +++ b/app/schemas/health.py @@ -108,3 +108,41 @@ class SystemStatusAdminResponse(BaseModel): default=None, description="One-click link to the backend CloudWatch log group, if configured", ) + + +class ComponentIncident(BaseModel): + """A single downtime period for one component (never ``healthy``).""" + + status: HealthStatus + startedAt: datetime + endedAt: datetime | None = Field( + default=None, description="Null while the incident is still ongoing" + ) + message: str | None = None + + +class ComponentDowntimeSummary(BaseModel): + """Downtime aggregates for a component over the requested time window.""" + + incidentCount: int + downtimeSeconds: float + uptimePercent: float = Field( + description="Share of the window the component was healthy, as a percentage" + ) + + +class ComponentDowntime(BaseModel): + """Downtime incidents and summary for a single component.""" + + name: str + summary: ComponentDowntimeSummary + incidents: list[ComponentIncident] + + +class SystemStatusDowntimeResponse(BaseModel): + """Verbose per-component downtime returned by GET /admin/api/system-status/history.""" + + windowHours: int + since: datetime + until: datetime + components: list[ComponentDowntime] diff --git a/app/services/health.py b/app/services/health.py index a8785a4..0e58123 100644 --- a/app/services/health.py +++ b/app/services/health.py @@ -23,10 +23,10 @@ from urllib.parse import quote import httpx -from sqlalchemy import func, select +from sqlalchemy import func, or_, select from sqlalchemy.orm import Session -from ..db.models.system_status import SystemStatusCache +from ..db.models.system_status import SystemStatusCache, SystemStatusIncident from ..schemas.health import ( COMPONENT_COMPUTE_ENV, COMPONENT_SEQERA_API, @@ -487,8 +487,57 @@ def _try_acquire_cache_refresh_lock(db: Session) -> bool: return bool(db.execute(select(func.pg_try_advisory_xact_lock(_CACHE_LOCK_ID))).scalar_one()) +def _open_incident(db: Session, status: SystemStatus, component: ProbeResult) -> None: + db.add( + SystemStatusIncident( + component=component.name, + status=component.status, + started_at=status.checked_at, + message=component.message, + ) + ) + + +def _update_incidents(db: Session, status: SystemStatus) -> None: + """Open/close/split incident rows so only downtime periods are stored. + + Healthy checks are never recorded. A component's most recent incident with + ``ended_at IS NULL`` (if any) is its currently-open outage: + + - Recovers to healthy -> close it. + - Still down at the same severity -> leave it open (refresh the message). + - Still down but severity changed (e.g. degraded -> unhealthy) -> close it + and open a new one, so each row represents a single severity level. + - No open incident and still down -> open a new one. + """ + for component in status.components: + open_incident = db.execute( + select(SystemStatusIncident) + .where(SystemStatusIncident.component == component.name) + .where(SystemStatusIncident.ended_at.is_(None)) + .order_by(SystemStatusIncident.started_at.desc()) + .limit(1) + ).scalar_one_or_none() + + if component.status == "healthy": + if open_incident is not None: + open_incident.ended_at = status.checked_at + db.add(open_incident) + continue + + if open_incident is None: + _open_incident(db, status, component) + elif open_incident.status != component.status: + open_incident.ended_at = status.checked_at + db.add(open_incident) + _open_incident(db, status, component) + elif component.message != open_incident.message: + open_incident.message = component.message + db.add(open_incident) + + async def refresh_db_cache(db: Session) -> SystemStatus: - """Run probes and replace the single shared DB cache row.""" + """Run probes, replace the single shared DB cache row, and log incidents.""" status = await collect_system_status() now = datetime.now(UTC) expires_at = now + timedelta(seconds=_CACHE_TTL_SECONDS) @@ -510,6 +559,7 @@ async def refresh_db_cache(db: Session) -> SystemStatus: row.updated_at = now db.add(row) + _update_incidents(db, status) db.commit() return status @@ -610,6 +660,95 @@ def to_public_dict(status: SystemStatus) -> dict[str, Any]: } +def _as_utc(value: datetime) -> datetime: + return value if value.tzinfo is not None else value.replace(tzinfo=UTC) + + +def get_incidents( + db: Session, + *, + since: datetime, + until: datetime, + component: str | None = None, +) -> list[SystemStatusIncident]: + """Return incidents overlapping ``[since, until]``, oldest first. + + An incident overlaps the window if it started at/before ``until`` and + either is still open (``ended_at IS NULL``) or ended at/after ``since``. + """ + stmt = select(SystemStatusIncident).where( + SystemStatusIncident.started_at <= until, + or_( + SystemStatusIncident.ended_at.is_(None), + SystemStatusIncident.ended_at >= since, + ), + ) + if component: + stmt = stmt.where(SystemStatusIncident.component == component) + stmt = stmt.order_by(SystemStatusIncident.component, SystemStatusIncident.started_at) + return list(db.execute(stmt).scalars()) + + +def _clipped_downtime_seconds( + incident: SystemStatusIncident, *, since: datetime, until: datetime +) -> float: + """Downtime contributed by ``incident`` within ``[since, until]``. + + Incidents that started before ``since`` or are still ongoing (``ended_at`` + is None, so it counts as down through ``until``) are clipped to the window + so a long-running outage doesn't over-count time outside the query range. + """ + start = max(_as_utc(incident.started_at), since) + end = min(_as_utc(incident.ended_at) if incident.ended_at is not None else until, until) + return max((end - start).total_seconds(), 0.0) + + +def _summarize_downtime( + incidents: list[SystemStatusIncident], *, since: datetime, until: datetime +) -> dict[str, Any]: + downtime_seconds = sum( + _clipped_downtime_seconds(incident, since=since, until=until) for incident in incidents + ) + window_seconds = max((until - since).total_seconds(), 0.0) + uptime_percent = ( + round(max(window_seconds - downtime_seconds, 0.0) / window_seconds * 100, 2) + if window_seconds + else 100.0 + ) + return { + "incidentCount": len(incidents), + "downtimeSeconds": round(downtime_seconds, 2), + "uptimePercent": uptime_percent, + } + + +def to_downtime_dict( + incidents: list[SystemStatusIncident], *, since: datetime, until: datetime, hours: int +) -> dict[str, Any]: + """Group flat incident rows by component into the admin downtime response shape.""" + by_component: dict[str, list[SystemStatusIncident]] = {} + for incident in incidents: + by_component.setdefault(incident.component, []).append(incident) + + components = [ + { + "name": name, + "summary": _summarize_downtime(rows, since=since, until=until), + "incidents": [ + { + "status": row.status, + "startedAt": row.started_at, + "endedAt": row.ended_at, + "message": row.message, + } + for row in rows + ], + } + for name, rows in by_component.items() + ] + return {"windowHours": hours, "since": since, "until": until, "components": components} + + def to_admin_dict(status: SystemStatus) -> dict[str, Any]: """Verbose projection: includes raw probe detail and a CloudWatch link.""" return { From 379a7f989cb50bd479fb744fe1712502f766dd79 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 24 Jul 2026 10:14:53 +1000 Subject: [PATCH 51/63] feat: display historic health status --- app/templates/admin/system_status.html | 201 ++++++++++++++++++++++ tests/test_routes_admin_system_status.py | 98 ++++++++++- tests/test_services_health.py | 202 ++++++++++++++++++++++- 3 files changed, 499 insertions(+), 2 deletions(-) diff --git a/app/templates/admin/system_status.html b/app/templates/admin/system_status.html index 7308a21..bc79c8c 100644 --- a/app/templates/admin/system_status.html +++ b/app/templates/admin/system_status.html @@ -56,6 +56,28 @@

Components

+
+
+
+

History

+
+ + + +
+
+
+
Loading…
+
+
+ + Only downtime is recorded: a colored segment marks a period the component was degraded (yellow) + or unhealthy (red). An all-green bar means no downtime in the selected window. + +
+
+
+ {% endblock %} diff --git a/tests/test_routes_admin_system_status.py b/tests/test_routes_admin_system_status.py index 1834af6..576310f 100644 --- a/tests/test_routes_admin_system_status.py +++ b/tests/test_routes_admin_system_status.py @@ -3,13 +3,15 @@ from __future__ import annotations import os -from datetime import UTC, datetime +from datetime import UTC, datetime, timedelta +import pytest from fastapi import FastAPI from fastapi.testclient import TestClient from pytest import MonkeyPatch from app.db.admin import require_admin_access +from app.db.models.system_status import SystemStatusIncident from app.routes.system_status import router as system_status_router from app.schemas.health import ProbeResult, SystemStatus from app.services import health @@ -136,3 +138,97 @@ def test_system_status_not_shadowed_by_admin_mount(mocker): assert resp.status_code == 401 # our endpoint's auth gate, not the admin mount assert "text/html" not in resp.headers.get("content-type", "") + + +# --------------------------------------------------------------------------- +# GET /admin/api/system-status/history (downtime incidents) +# --------------------------------------------------------------------------- + + +def _seed_incident(test_db, *, component, status, hours_ago, ended_hours_ago=None, message=None): + test_db.add( + SystemStatusIncident( + component=component, + status=status, + started_at=datetime.now(UTC) - timedelta(hours=hours_ago), + ended_at=( + datetime.now(UTC) - timedelta(hours=ended_hours_ago) + if ended_hours_ago is not None + else None + ), + message=message, + ) + ) + test_db.commit() + + +def test_history_requires_admin(client): + resp = client.get("/admin/api/system-status/history") + assert resp.status_code == 401 + + +def test_history_returns_per_component_summary_and_incidents(client, test_db): + client.app.dependency_overrides[require_admin_access] = lambda: {"sub": "auth0|admin"} + _seed_incident( + test_db, + component="seqera_compute_env", + status="unhealthy", + hours_ago=2, + ended_hours_ago=1, + message="ERRORED", + ) + + resp = client.get("/admin/api/system-status/history?hours=24") + assert resp.status_code == 200 + body = resp.json() + assert body["windowHours"] == 24 + assert "since" in body and "until" in body + + by_name = {c["name"]: c for c in body["components"]} + ce = by_name["seqera_compute_env"] + assert ce["summary"]["incidentCount"] == 1 + assert ce["summary"]["downtimeSeconds"] == pytest.approx(3600, rel=0.01) + assert ce["incidents"][0]["message"] == "ERRORED" + assert ce["incidents"][0]["endedAt"] is not None + + +def test_history_includes_ongoing_incident(client, test_db): + client.app.dependency_overrides[require_admin_access] = lambda: {"sub": "auth0|admin"} + _seed_incident(test_db, component="seqera_api", status="unhealthy", hours_ago=1) + + resp = client.get("/admin/api/system-status/history?hours=24") + assert resp.status_code == 200 + body = resp.json() + api = next(c for c in body["components"] if c["name"] == "seqera_api") + assert api["incidents"][0]["endedAt"] is None + assert api["summary"]["incidentCount"] == 1 + + +def test_history_excludes_incidents_outside_window(client, test_db): + client.app.dependency_overrides[require_admin_access] = lambda: {"sub": "auth0|admin"} + _seed_incident( + test_db, component="seqera_api", status="unhealthy", hours_ago=48, ended_hours_ago=47 + ) + + resp = client.get("/admin/api/system-status/history?hours=24") + assert resp.status_code == 200 + assert resp.json()["components"] == [] + + +def test_history_filters_by_component_query_param(client, test_db): + client.app.dependency_overrides[require_admin_access] = lambda: {"sub": "auth0|admin"} + _seed_incident(test_db, component="seqera_api", status="unhealthy", hours_ago=1) + _seed_incident(test_db, component="seqera_compute_env", status="unhealthy", hours_ago=1) + + resp = client.get("/admin/api/system-status/history?hours=24&component=seqera_api") + assert resp.status_code == 200 + body = resp.json() + assert len(body["components"]) == 1 + assert body["components"][0]["name"] == "seqera_api" + + +def test_history_empty_when_no_incidents_recorded(client): + client.app.dependency_overrides[require_admin_access] = lambda: {"sub": "auth0|admin"} + resp = client.get("/admin/api/system-status/history") + assert resp.status_code == 200 + assert resp.json()["components"] == [] diff --git a/tests/test_services_health.py b/tests/test_services_health.py index 47a6d20..5aebb01 100644 --- a/tests/test_services_health.py +++ b/tests/test_services_health.py @@ -5,8 +5,9 @@ from datetime import UTC, datetime, timedelta import httpx +from sqlalchemy import select -from app.db.models.system_status import SystemStatusCache +from app.db.models.system_status import SystemStatusCache, SystemStatusIncident from app.schemas.health import ( COMPONENT_COMPUTE_ENV, COMPONENT_SEQERA_API, @@ -216,6 +217,205 @@ async def ok_get(self, url, *args, **kwargs): # noqa: ANN001 assert stale.checked_at == first.checked_at +# --------------------------------------------------------------------------- +# Downtime incidents (only degraded/unhealthy periods are recorded) +# --------------------------------------------------------------------------- + + +def _incidents(db, component=None): + rows = db.execute(select(SystemStatusIncident)).scalars().all() + return [r for r in rows if component is None or r.component == component] + + +async def test_healthy_refresh_writes_no_incidents(monkeypatch, test_db): + _mock_response( + monkeypatch, + user_info=_ok_user_info, + compute_env=_compute_env_with_status("AVAILABLE"), + ) + await health.get_system_status(test_db, force_refresh=True) + + assert _incidents(test_db) == [] + + +async def test_going_unhealthy_opens_an_incident(monkeypatch, test_db): + _mock_response( + monkeypatch, + user_info=_ok_user_info, + compute_env=_compute_env_with_status("ERRORED", "Gadi agent disconnected"), + ) + status = await health.get_system_status(test_db, force_refresh=True) + + rows = _incidents(test_db, COMPONENT_COMPUTE_ENV) + assert len(rows) == 1 + assert rows[0].status == "unhealthy" + assert rows[0].ended_at is None + assert "Gadi agent disconnected" in (rows[0].message or "") + assert rows[0].started_at == status.checked_at.replace(tzinfo=None) + + +async def test_recovering_to_healthy_closes_the_incident(monkeypatch, test_db): + _mock_response( + monkeypatch, + user_info=_ok_user_info, + compute_env=_compute_env_with_status("ERRORED"), + ) + await health.get_system_status(test_db, force_refresh=True) + + _mock_response( + monkeypatch, + user_info=_ok_user_info, + compute_env=_compute_env_with_status("AVAILABLE"), + ) + recovered = await health.get_system_status(test_db, force_refresh=True) + + rows = _incidents(test_db, COMPONENT_COMPUTE_ENV) + assert len(rows) == 1 # closed, not duplicated + assert rows[0].ended_at == recovered.checked_at.replace(tzinfo=None) + + +async def test_staying_unhealthy_does_not_open_a_second_incident(monkeypatch, test_db): + _mock_response( + monkeypatch, + user_info=_ok_user_info, + compute_env=_compute_env_with_status("ERRORED", "first error"), + ) + await health.get_system_status(test_db, force_refresh=True) + + _mock_response( + monkeypatch, + user_info=_ok_user_info, + compute_env=_compute_env_with_status("ERRORED", "still erroring"), + ) + await health.get_system_status(test_db, force_refresh=True) + + rows = _incidents(test_db, COMPONENT_COMPUTE_ENV) + assert len(rows) == 1 + assert rows[0].ended_at is None + # The message on the still-open incident is refreshed to the latest probe result. + assert rows[0].message == "Compute environment state: ERRORED (still erroring)" + + +async def test_severity_change_splits_the_incident(monkeypatch, test_db): + _mock_response( + monkeypatch, + user_info=_ok_user_info, + compute_env=_compute_env_with_status("CREATING"), # degraded + ) + await health.get_system_status(test_db, force_refresh=True) + + _mock_response( + monkeypatch, + user_info=_ok_user_info, + compute_env=_compute_env_with_status("ERRORED"), # unhealthy + ) + await health.get_system_status(test_db, force_refresh=True) + + rows = sorted( + _incidents(test_db, COMPONENT_COMPUTE_ENV), key=lambda r: r.started_at + ) + assert len(rows) == 2 + assert rows[0].status == "degraded" + assert rows[0].ended_at is not None # closed when severity changed + assert rows[1].status == "unhealthy" + assert rows[1].ended_at is None + + +def test_get_incidents_filters_by_window_and_component(test_db): + now = datetime.now(UTC) + long_ago = SystemStatusIncident( + component=COMPONENT_SEQERA_API, + status="unhealthy", + started_at=now - timedelta(hours=5), + ended_at=now - timedelta(hours=4, minutes=55), + ) + recent_api = SystemStatusIncident( + component=COMPONENT_SEQERA_API, + status="unhealthy", + started_at=now - timedelta(minutes=30), + ended_at=now - timedelta(minutes=25), + ) + ongoing_ce = SystemStatusIncident( + component=COMPONENT_COMPUTE_ENV, + status="degraded", + started_at=now - timedelta(minutes=10), + ended_at=None, + ) + test_db.add_all([long_ago, recent_api, ongoing_ce]) + test_db.commit() + + since = now - timedelta(hours=1) + all_recent = health.get_incidents(test_db, since=since, until=now) + assert {row.component for row in all_recent} == {COMPONENT_SEQERA_API, COMPONENT_COMPUTE_ENV} + + api_only = health.get_incidents( + test_db, since=since, until=now, component=COMPONENT_SEQERA_API + ) + assert len(api_only) == 1 + assert api_only[0].ended_at is not None + + +def test_summarize_downtime_clips_to_window(): + now = datetime.now(UTC) + since = now - timedelta(hours=1) + incidents = [ + # Started before the window, ended inside it -> clipped at `since`. + SystemStatusIncident( + component="x", + status="unhealthy", + started_at=now - timedelta(hours=2), + ended_at=since + timedelta(minutes=10), + ), + # Still ongoing -> counts through `until`. + SystemStatusIncident( + component="x", + status="degraded", + started_at=now - timedelta(minutes=5), + ended_at=None, + ), + ] + summary = health._summarize_downtime(incidents, since=since, until=now) + + assert summary["incidentCount"] == 2 + assert summary["downtimeSeconds"] == 10 * 60 + 5 * 60 + assert summary["uptimePercent"] == round((3600 - 900) / 3600 * 100, 2) + + +def test_summarize_downtime_no_incidents_is_full_uptime(): + now = datetime.now(UTC) + summary = health._summarize_downtime([], since=now - timedelta(hours=1), until=now) + assert summary == {"incidentCount": 0, "downtimeSeconds": 0.0, "uptimePercent": 100.0} + + +def test_to_downtime_dict_groups_by_component(): + now = datetime.now(UTC) + since = now - timedelta(hours=24) + incidents = [ + SystemStatusIncident( + component=COMPONENT_SEQERA_API, + status="unhealthy", + started_at=since + timedelta(hours=1), + ended_at=since + timedelta(hours=2), + ), + SystemStatusIncident( + component=COMPONENT_COMPUTE_ENV, + status="degraded", + started_at=since + timedelta(hours=3), + ended_at=None, + message="still starting up", + ), + ] + result = health.to_downtime_dict(incidents, since=since, until=now, hours=24) + + assert result["windowHours"] == 24 + assert result["since"] == since + assert result["until"] == now + by_name = {c["name"]: c for c in result["components"]} + assert by_name[COMPONENT_SEQERA_API]["summary"]["incidentCount"] == 1 + assert by_name[COMPONENT_SEQERA_API]["incidents"][0]["endedAt"] == since + timedelta(hours=2) + assert by_name[COMPONENT_COMPUTE_ENV]["incidents"][0]["message"] == "still starting up" + + def test_overall_status_aggregation(): assert health._worst(["healthy", "healthy"]) == "healthy" assert health._worst(["healthy", "degraded"]) == "degraded" From 484788b8b769ed58ada679143c6693b0ea0df9bb Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Fri, 24 Jul 2026 13:06:31 +1000 Subject: [PATCH 52/63] chore: lint --- app/routes/system_status.py | 4 +--- tests/test_services_health.py | 8 ++------ 2 files changed, 3 insertions(+), 9 deletions(-) diff --git a/app/routes/system_status.py b/app/routes/system_status.py index 1fe478f..fc6b9d0 100644 --- a/app/routes/system_status.py +++ b/app/routes/system_status.py @@ -52,9 +52,7 @@ async def get_admin_system_status_history( le=_MAX_HISTORY_HOURS, description="How many hours of downtime history to return", ), - component: str | None = Query( - default=None, description="Restrict to a single component name" - ), + component: str | None = Query(default=None, description="Restrict to a single component name"), db: Session = Depends(get_db), ) -> SystemStatusDowntimeResponse: """Return per-component downtime incidents, so admins can see how frequently diff --git a/tests/test_services_health.py b/tests/test_services_health.py index 5aebb01..322f2bf 100644 --- a/tests/test_services_health.py +++ b/tests/test_services_health.py @@ -311,9 +311,7 @@ async def test_severity_change_splits_the_incident(monkeypatch, test_db): ) await health.get_system_status(test_db, force_refresh=True) - rows = sorted( - _incidents(test_db, COMPONENT_COMPUTE_ENV), key=lambda r: r.started_at - ) + rows = sorted(_incidents(test_db, COMPONENT_COMPUTE_ENV), key=lambda r: r.started_at) assert len(rows) == 2 assert rows[0].status == "degraded" assert rows[0].ended_at is not None # closed when severity changed @@ -348,9 +346,7 @@ def test_get_incidents_filters_by_window_and_component(test_db): all_recent = health.get_incidents(test_db, since=since, until=now) assert {row.component for row in all_recent} == {COMPONENT_SEQERA_API, COMPONENT_COMPUTE_ENV} - api_only = health.get_incidents( - test_db, since=since, until=now, component=COMPONENT_SEQERA_API - ) + api_only = health.get_incidents(test_db, since=since, until=now, component=COMPONENT_SEQERA_API) assert len(api_only) == 1 assert api_only[0].ended_at is not None From a26dd97cbb5493e4d2b585bf9ee470aeff885920 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Mon, 27 Jul 2026 09:25:11 +1000 Subject: [PATCH 53/63] feat: fix summary time min width --- app/templates/admin/system_status.html | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/app/templates/admin/system_status.html b/app/templates/admin/system_status.html index bc79c8c..b38951e 100644 --- a/app/templates/admin/system_status.html +++ b/app/templates/admin/system_status.html @@ -387,11 +387,12 @@

History

topRow.className = "d-flex align-items-center"; var nameCol = document.createElement("div"); - nameCol.style.width = "220px"; + nameCol.style.minWidth = "220px"; nameCol.style.flexShrink = "0"; + nameCol.style.marginRight = "1rem"; nameCol.innerHTML = "" + escapeHtml(component.name) + "" + infoIcon(component.name) + - '
' + + '
' + summary.uptimePercent + "% uptime · " + summary.incidentCount + " incident" + (summary.incidentCount === 1 ? "" : "s") + (summary.downtimeSeconds > 0 ? " · " + formatDuration(summary.downtimeSeconds) + " down" : "") + From 33627096d20c2f62317b05999d78693b14ad18b7 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Mon, 27 Jul 2026 09:33:56 +1000 Subject: [PATCH 54/63] refactor: system status template --- app/templates/admin/system_status.html | 51 +++++++++++--------------- tests/test_services_health.py | 15 +++++++- 2 files changed, 35 insertions(+), 31 deletions(-) diff --git a/app/templates/admin/system_status.html b/app/templates/admin/system_status.html index b38951e..1174fa9 100644 --- a/app/templates/admin/system_status.html +++ b/app/templates/admin/system_status.html @@ -11,6 +11,12 @@

System Status

{% endblock %} {% block content %} +
@@ -33,7 +39,7 @@

Components

Status Latency - @@ -143,15 +149,18 @@

History

return base + note; } + // Common `data-bs-toggle="tooltip" ... title="..."` attribute string, shared by + // every element below that shows a hover tooltip. + function tooltipAttrs(title) { + if (!title) return ""; + return ' data-bs-toggle="tooltip" data-bs-placement="top" title="' + escapeHtml(title) + '"'; + } + function latencyCell(component) { var ms = component.latencyMs; if (ms === null || ms === undefined) return "—"; return ( - '' + - ms + - " ms" + '" + ms + " ms" ); } @@ -159,10 +168,7 @@

History

var desc = DESCRIPTIONS[name]; if (!desc) return ""; return ( - ' ' + ' " ); } @@ -191,12 +197,8 @@

History

var cls = pillClass[status] || DEFAULT_PILL; var label = status ? status.charAt(0).toUpperCase() + status.slice(1) : "Unknown"; var desc = STATUS_DESCRIPTIONS[status]; - var tip = desc - ? ' style="cursor:help" data-bs-toggle="tooltip" data-bs-placement="top" title="' + - escapeHtml(desc) + - '"' - : ""; - return '" + label + ""; + var helpClass = desc ? " help-cursor" : ""; + return '" + label + ""; } function escapeHtml(value) { @@ -306,12 +308,7 @@

History

function incidentBar(incidents, sinceMs, untilMs) { var windowMs = Math.max(untilMs - sinceMs, 1); var bar = document.createElement("div"); - bar.className = "bg-success-lt"; - bar.style.position = "relative"; - bar.style.height = "20px"; - bar.style.borderRadius = "4px"; - bar.style.flex = "1"; - bar.style.overflow = "hidden"; + bar.className = "bg-success-lt status-bar"; incidents.forEach(function (incident) { var startMs = Math.max(new Date(incident.startedAt).getTime(), sinceMs); @@ -322,13 +319,9 @@

History

if (endMs <= startMs) return; var seg = document.createElement("div"); - seg.className = incident.status === "unhealthy" ? "bg-danger" : "bg-warning"; - seg.style.position = "absolute"; - seg.style.top = "0"; - seg.style.bottom = "0"; + seg.className = (incident.status === "unhealthy" ? "bg-danger" : "bg-warning") + " status-bar-segment"; seg.style.left = ((startMs - sinceMs) / windowMs) * 100 + "%"; seg.style.width = Math.max(((endMs - startMs) / windowMs) * 100, 0.5) + "%"; - seg.style.cursor = "help"; seg.setAttribute("data-bs-toggle", "tooltip"); seg.setAttribute("data-bs-placement", "top"); var label = @@ -387,9 +380,7 @@

History

topRow.className = "d-flex align-items-center"; var nameCol = document.createElement("div"); - nameCol.style.minWidth = "220px"; - nameCol.style.flexShrink = "0"; - nameCol.style.marginRight = "1rem"; + nameCol.className = "history-name-col"; nameCol.innerHTML = "" + escapeHtml(component.name) + "" + infoIcon(component.name) + '
' + diff --git a/tests/test_services_health.py b/tests/test_services_health.py index 322f2bf..cff67e4 100644 --- a/tests/test_services_health.py +++ b/tests/test_services_health.py @@ -5,6 +5,7 @@ from datetime import UTC, datetime, timedelta import httpx +import pytest from sqlalchemy import select from app.db.models.system_status import SystemStatusCache, SystemStatusIncident @@ -16,6 +17,19 @@ from app.services import health +@pytest.fixture(autouse=True) +def _agent_healthcheck_disabled_by_default(monkeypatch): + """Keep these tests hermetic regardless of a developer's local .env. + + The Tower Agent probe is opt-in via ENABLE_AGENT_HEALTHCHECK, and app.main + loads .env at import time, so a developer with it set locally would + otherwise get an unmocked probe_tower_agent() running in every test here. + Tests that specifically exercise the agent probe re-enable it themselves + (see _install_agent_mock). + """ + monkeypatch.delenv("ENABLE_AGENT_HEALTHCHECK", raising=False) + + def _component(status: health.SystemStatus, name: str) -> health.ProbeResult: return next(c for c in status.components if c.name == name) @@ -517,7 +531,6 @@ async def fake_delete(self, url, *args, **kwargs): # noqa: ANN001 async def test_agent_probe_disabled_by_default(monkeypatch): - monkeypatch.delenv("ENABLE_AGENT_HEALTHCHECK", raising=False) _mock_response( monkeypatch, user_info=_ok_user_info, From 65023eccaf22257c4757d815596e8fbf9acde494 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Mon, 27 Jul 2026 10:44:35 +1000 Subject: [PATCH 55/63] feat: detect time zone in system status --- app/templates/admin/system_status.html | 50 ++++++++++++++++++++++---- 1 file changed, 44 insertions(+), 6 deletions(-) diff --git a/app/templates/admin/system_status.html b/app/templates/admin/system_status.html index 1174fa9..e790cdc 100644 --- a/app/templates/admin/system_status.html +++ b/app/templates/admin/system_status.html @@ -52,7 +52,9 @@

Components

- Auto-refreshes every 30s. + + Auto-refreshes every 30s. Times shown in . +
@@ -90,6 +93,33 @@

History

var ENDPOINT = "/admin/api/system-status"; var REFRESH_MS = 30000; + // All timestamps below are rendered with toLocaleString(), i.e. in whatever + // timezone the viewer's own browser/OS is set to (not a fixed timezone like + // the rest of the admin dashboard). Label it explicitly so it's unambiguous + // when cross-referencing against UTC-timestamped CloudWatch logs. + var DETECTED_TZ = (function () { + try { + return Intl.DateTimeFormat().resolvedOptions().timeZone || "your browser's local time"; + } catch (e) { + return "your browser's local time"; + } + })(); + document.getElementById("components-tz").textContent = DETECTED_TZ; + document.getElementById("history-tz").textContent = DETECTED_TZ; + + // Short zone abbreviation (e.g. "AEST"/"AEDT") for a given instant, so each + // individual timestamp is unambiguous without having to check the footer. + // Computed per-date (not just once) since it can change across a DST boundary. + function tzAbbrev(date) { + try { + var parts = Intl.DateTimeFormat(undefined, { timeZoneName: "short" }).formatToParts(date); + var part = parts.find(function (p) { return p.type === "timeZoneName"; }); + return part ? part.value : ""; + } catch (e) { + return ""; + } + } + // text-white + fw-semibold keeps the label readable on the colored fill // (Tabler badges default to a muted text color, which washed out on green). var pillClass = { @@ -232,7 +262,12 @@

History

: "Unknown"; var checkedAt = document.getElementById("checked-at"); - checkedAt.textContent = data.checkedAt ? new Date(data.checkedAt).toLocaleString() : "—"; + if (data.checkedAt) { + var checkedAtDate = new Date(data.checkedAt); + checkedAt.textContent = checkedAtDate.toLocaleString() + " " + tzAbbrev(checkedAtDate); + } else { + checkedAt.textContent = "—"; + } var body = document.getElementById("components-body"); disposeTooltips(body); // tear down tooltips on the rows we're replacing @@ -324,10 +359,12 @@

History

seg.style.width = Math.max(((endMs - startMs) / windowMs) * 100, 0.5) + "%"; seg.setAttribute("data-bs-toggle", "tooltip"); seg.setAttribute("data-bs-placement", "top"); + var incidentStart = new Date(incident.startedAt); var label = (incident.status === "unhealthy" ? "Unhealthy" : "Degraded") + - ": " + new Date(incident.startedAt).toLocaleString() + + ": " + incidentStart.toLocaleString() + " → " + (incident.endedAt ? new Date(incident.endedAt).toLocaleString() : "ongoing") + + " " + tzAbbrev(incidentStart) + (incident.message ? " (" + incident.message + ")" : ""); seg.setAttribute("title", label); bar.appendChild(seg); @@ -337,17 +374,18 @@

History

} function incidentListItem(incident) { + var startDate = new Date(incident.startedAt); var durationSeconds = - (incident.endedAt ? new Date(incident.endedAt).getTime() : Date.now()) - - new Date(incident.startedAt).getTime(); + (incident.endedAt ? new Date(incident.endedAt).getTime() : Date.now()) - startDate.getTime(); var li = document.createElement("li"); li.className = "small"; li.innerHTML = badge(incident.status) + " " + - escapeHtml(new Date(incident.startedAt).toLocaleString()) + + escapeHtml(startDate.toLocaleString()) + " → " + (incident.endedAt ? escapeHtml(new Date(incident.endedAt).toLocaleString()) : 'ongoing') + + ' ' + escapeHtml(tzAbbrev(startDate)) + "" + ' (' + formatDuration(durationSeconds / 1000) + ")" + (incident.message ? ' — ' + escapeHtml(incident.message) + "" From 32eaf54b0d8443b06a008ab647849d33313de948 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Mon, 27 Jul 2026 13:37:56 +1000 Subject: [PATCH 56/63] fix: fix out_dir in proteindj workflow --- app/services/proteindj_config.py | 2 +- tests/test_proteindj_coverage.py | 6 +++--- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/app/services/proteindj_config.py b/app/services/proteindj_config.py index 1f73557..732711c 100644 --- a/app/services/proteindj_config.py +++ b/app/services/proteindj_config.py @@ -21,7 +21,7 @@ def get_proteindj_default_params( no samplesheet — so these are passed straight through as paramsText keys. """ return { - "outdir": out_dir, + "out_dir": out_dir, "input_pdb": input_pdb, "hotspot_residues": hotspot_residues, "num_designs": num_designs, diff --git a/tests/test_proteindj_coverage.py b/tests/test_proteindj_coverage.py index 1c863b1..7afd0b9 100644 --- a/tests/test_proteindj_coverage.py +++ b/tests/test_proteindj_coverage.py @@ -69,7 +69,7 @@ def _queued_proteindj_job( "workspaceId": "ws_123", "revision": "dev", "paramsText": params_text - or ("outdir: s3://my-bucket/run-output-id\ninput_pdb: s3://my-bucket/inputs/test.pdb"), + or ("out_dir: s3://my-bucket/run-output-id\ninput_pdb: s3://my-bucket/inputs/test.pdb"), "configProfiles": ["singularity"], "configText": "config_text", "resume": False, @@ -182,7 +182,7 @@ def test_get_proteindj_default_params_all_fields(): design_length="100-150", ) assert params == { - "outdir": "s3://bucket/out", + "out_dir": "s3://bucket/out", "input_pdb": "s3://bucket/in.pdb", "hotspot_residues": "A20,A21", "num_designs": 5, @@ -328,7 +328,7 @@ async def test_prepare_proteindj_workflow_writes_expected_queued_job( assert "preRunScript" not in queued_job.launch_payload assert queued_job.launch_payload["resume"] is False params_text = queued_job.launch_payload["paramsText"] - assert "outdir: s3://my-bucket/run-output-id" in params_text + assert "out_dir: s3://my-bucket/run-output-id" in params_text assert "input_pdb: s3://my-bucket/inputs/test.pdb" in params_text assert "hotspot_residues: A20,A21" in params_text assert "num_designs: 5" in params_text From cea3758a0b3338eea9431dce1d8290661f4922a3 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Mon, 27 Jul 2026 16:03:44 +1000 Subject: [PATCH 57/63] feat: rfdiffusion results --- app/services/results_utils.py | 42 +++++++++++++++++++++++++++++++++-- 1 file changed, 40 insertions(+), 2 deletions(-) diff --git a/app/services/results_utils.py b/app/services/results_utils.py index 93c24e4..abc3a00 100644 --- a/app/services/results_utils.py +++ b/app/services/results_utils.py @@ -173,7 +173,7 @@ async def resolve_fasta_form_data( expiration=3600, response_content_disposition=_format_attachment_content_disposition(filename), ) - except S3ConfigurationError, S3ServiceError: + except (S3ConfigurationError, S3ServiceError): logger.warning( "Failed to generate presigned URL for %r (S3 key %r)", key, @@ -213,7 +213,7 @@ async def resolve_pdb_presigned_urls( response_content_disposition=_format_attachment_content_disposition(filename), ) return {**form_data, "starting_pdb": presigned_url} - except S3ConfigurationError, S3ServiceError: + except (S3ConfigurationError, S3ServiceError): logger.warning( "Failed to generate presigned starting_pdb URL for S3 key %r; " "returning original form data", @@ -654,6 +654,35 @@ def build_colabfold_proteinfold_output_listing_prefixes(run: WorkflowRun) -> lis return prefixes +def classify_rfdiffusion_output_key( + key: str, sample_id: str | None = None +) -> ClassifiedOutput | None: + normalized = key.strip() + if not normalized or normalized.endswith("/"): + return None + basename = normalized.rsplit("/", 1)[-1] + if "/results/" in normalized.lower() and basename == "ranked_designs.csv": + return ClassifiedOutput(category="stats_csv", label=basename) + return None + + +def get_rfdiffusion_score_file(keys: list[str], sample_id: str | None) -> str | None: + return None + + +async def extract_rfdiffusion_max_score(score_file: str) -> float | None: + # get_rfdiffusion_score_file always returns None (ranked_designs.csv's score + # column isn't documented yet), so this is never actually invoked. + return None + + +def build_rfdiffusion_output_listing_prefixes(run: WorkflowRun) -> list[str]: + run_uuid = str(getattr(run, "id", "") or "").strip() + if not run_uuid: + return [] + return [f"{run_uuid}/results/"] + + def _make_wisps_spec(tool: WorkflowTool) -> WorkflowResultsSpec: return WorkflowResultsSpec( kind="interaction-screening", @@ -690,6 +719,15 @@ def _make_bulk_prediction_spec(tool: WorkflowTool) -> WorkflowResultsSpec: classify=classify_bindcraft_output_key, supports_snapshots=True, ), + "rfdiffusion": WorkflowResultsSpec( + kind="de-novo-design", + tool="rfdiffusion", + required_categories={"stats_csv"}, + get_prefixes=build_rfdiffusion_output_listing_prefixes, + get_score_file=get_rfdiffusion_score_file, + extract_max_score=extract_rfdiffusion_max_score, + classify=classify_rfdiffusion_output_key, + ), }, "single-prediction": { "boltz": WorkflowResultsSpec( From a3536fbb45bed6250c0f240284f0780f42e6c8d0 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Mon, 27 Jul 2026 16:08:52 +1000 Subject: [PATCH 58/63] feat: update nextflow module version --- app/services/bindflow_executor.py | 9 +++++++-- app/services/launch_payloads.py | 2 ++ app/services/proteindj_executor.py | 9 +++++++-- app/services/proteinfold_executor.py | 9 +++++++-- 4 files changed, 23 insertions(+), 6 deletions(-) diff --git a/app/services/bindflow_executor.py b/app/services/bindflow_executor.py index a29ca05..7f75a92 100644 --- a/app/services/bindflow_executor.py +++ b/app/services/bindflow_executor.py @@ -16,7 +16,12 @@ get_bindflow_config_text, get_bindflow_default_params, ) -from .launch_payloads import get_executor_script, inject_prerun_script, without_prerun_script +from .launch_payloads import ( + DEFAULT_MODULE_LOADS, + get_executor_script, + inject_prerun_script, + without_prerun_script, +) from .seqera import ( WorkflowLaunchResult, _get_required_env, @@ -129,7 +134,7 @@ async def launch_bindflow_workflow( # pylint: disable=too-many-locals prerun_script = get_executor_script( prerun_script_path=queued_job.workflow.prerun_script_path, - module_loads=["singularity", "nextflow"], + module_loads=DEFAULT_MODULE_LOADS, env={ "AWS_ACCESS_KEY_ID": os.getenv("AWS_ACCESS_KEY_ID", ""), "AWS_SECRET_ACCESS_KEY": os.getenv("AWS_SECRET_ACCESS_KEY", ""), diff --git a/app/services/launch_payloads.py b/app/services/launch_payloads.py index 130ff7a..6efdfeb 100644 --- a/app/services/launch_payloads.py +++ b/app/services/launch_payloads.py @@ -7,6 +7,8 @@ from .workflow_config_fetcher import fetch_workflow_config +DEFAULT_MODULE_LOADS = ["singularity", "nextflow/25.10.3"] + def without_prerun_script(launch_payload: dict[str, Any]) -> dict[str, Any]: """Return a copy safe to persist in the job queue.""" diff --git a/app/services/proteindj_executor.py b/app/services/proteindj_executor.py index 45309ce..345d073 100644 --- a/app/services/proteindj_executor.py +++ b/app/services/proteindj_executor.py @@ -17,7 +17,12 @@ WorkflowLaunchForm, WorkflowUserDetails, ) -from .launch_payloads import get_executor_script, inject_prerun_script, without_prerun_script +from .launch_payloads import ( + DEFAULT_MODULE_LOADS, + get_executor_script, + inject_prerun_script, + without_prerun_script, +) from .proteindj_config import ( get_proteindj_config_profiles, get_proteindj_config_text, @@ -155,7 +160,7 @@ async def launch_proteindj_workflow( # pylint: disable=too-many-locals prerun_script = get_executor_script( prerun_script_path=queued_job.workflow.prerun_script_path, - module_loads=["singularity", "nextflow"], + module_loads=DEFAULT_MODULE_LOADS, env={ "AWS_ACCESS_KEY_ID": os.getenv("AWS_ACCESS_KEY_ID", ""), "AWS_SECRET_ACCESS_KEY": os.getenv("AWS_SECRET_ACCESS_KEY", ""), diff --git a/app/services/proteinfold_executor.py b/app/services/proteinfold_executor.py index bf73a50..55886f0 100644 --- a/app/services/proteinfold_executor.py +++ b/app/services/proteinfold_executor.py @@ -11,7 +11,12 @@ from ..db.models import QueuedJob, WorkflowRun from ..schemas.workflows import WorkflowFormData, WorkflowLaunchForm, WorkflowUserDetails -from .launch_payloads import get_executor_script, inject_prerun_script, without_prerun_script +from .launch_payloads import ( + DEFAULT_MODULE_LOADS, + get_executor_script, + inject_prerun_script, + without_prerun_script, +) from .proteinfold_config import ( get_proteinfold_config_profiles, get_proteinfold_config_text, @@ -149,7 +154,7 @@ async def launch_proteinfold_workflow( prerun_script = get_executor_script( prerun_script_path=queued_job.workflow.prerun_script_path, - module_loads=["singularity", "nextflow"], + module_loads=DEFAULT_MODULE_LOADS, env={ "AWS_ACCESS_KEY_ID": os.getenv("AWS_ACCESS_KEY_ID", ""), "AWS_SECRET_ACCESS_KEY": os.getenv("AWS_SECRET_ACCESS_KEY", ""), From 4a1f521eb5e2b9edc570526a6962c8895c5a74fa Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Mon, 27 Jul 2026 16:34:46 +1000 Subject: [PATCH 59/63] fix: update to nextflow 25.10.4 --- app/services/launch_payloads.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/app/services/launch_payloads.py b/app/services/launch_payloads.py index 6efdfeb..6bdeb79 100644 --- a/app/services/launch_payloads.py +++ b/app/services/launch_payloads.py @@ -7,7 +7,7 @@ from .workflow_config_fetcher import fetch_workflow_config -DEFAULT_MODULE_LOADS = ["singularity", "nextflow/25.10.3"] +DEFAULT_MODULE_LOADS = ["singularity", "nextflow/25.10.4"] def without_prerun_script(launch_payload: dict[str, Any]) -> dict[str, Any]: From a07b0687a589bf522cc52fc392cda5135a0eef72 Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Tue, 28 Jul 2026 10:51:38 +1000 Subject: [PATCH 60/63] feat: max score, archive pdb file for rfdiffusion --- app/services/results_utils.py | 25 +++++++++++++++++++++---- 1 file changed, 21 insertions(+), 4 deletions(-) diff --git a/app/services/results_utils.py b/app/services/results_utils.py index abc3a00..f22a2bb 100644 --- a/app/services/results_utils.py +++ b/app/services/results_utils.py @@ -661,18 +661,35 @@ def classify_rfdiffusion_output_key( if not normalized or normalized.endswith("/"): return None basename = normalized.rsplit("/", 1)[-1] - if "/results/" in normalized.lower() and basename == "ranked_designs.csv": + if "/results/" not in normalized.lower(): + return None + if basename == "ranked_designs.csv": return ClassifiedOutput(category="stats_csv", label=basename) + if basename == "ranked_designs.tar.gz": + return ClassifiedOutput(category="pdb", label=basename) return None def get_rfdiffusion_score_file(keys: list[str], sample_id: str | None) -> str | None: + for key in keys: + normalized = key.strip() + if not normalized: + continue + basename = normalized.rsplit("/", 1)[-1] + if basename == "ranked_designs.csv": + return normalized return None async def extract_rfdiffusion_max_score(score_file: str) -> float | None: - # get_rfdiffusion_score_file always returns None (ranked_designs.csv's score - # column isn't documented yet), so this is never actually invoked. + content = await read_s3_file(score_file) + csv_reader = csv.DictReader(StringIO(content)) + row = next(csv_reader, None) + if row is None: + return None + value = row.get("af2_plddt_overall") + if value and value.strip(): + return float(value) / 100 return None @@ -722,7 +739,7 @@ def _make_bulk_prediction_spec(tool: WorkflowTool) -> WorkflowResultsSpec: "rfdiffusion": WorkflowResultsSpec( kind="de-novo-design", tool="rfdiffusion", - required_categories={"stats_csv"}, + required_categories={"stats_csv", "pdb"}, get_prefixes=build_rfdiffusion_output_listing_prefixes, get_score_file=get_rfdiffusion_score_file, extract_max_score=extract_rfdiffusion_max_score, From cacac0870c2c4c293d5ce057b6a91dcdcee44f8c Mon Sep 17 00:00:00 2001 From: Anne Phan <53896516+vtnphan@users.noreply.github.com> Date: Tue, 28 Jul 2026 11:29:05 +1000 Subject: [PATCH 61/63] chore: lint --- app/services/results_utils.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/app/services/results_utils.py b/app/services/results_utils.py index f22a2bb..42eb2f7 100644 --- a/app/services/results_utils.py +++ b/app/services/results_utils.py @@ -173,7 +173,7 @@ async def resolve_fasta_form_data( expiration=3600, response_content_disposition=_format_attachment_content_disposition(filename), ) - except (S3ConfigurationError, S3ServiceError): + except S3ConfigurationError, S3ServiceError: logger.warning( "Failed to generate presigned URL for %r (S3 key %r)", key, @@ -213,7 +213,7 @@ async def resolve_pdb_presigned_urls( response_content_disposition=_format_attachment_content_disposition(filename), ) return {**form_data, "starting_pdb": presigned_url} - except (S3ConfigurationError, S3ServiceError): + except S3ConfigurationError, S3ServiceError: logger.warning( "Failed to generate presigned starting_pdb URL for S3 key %r; " "returning original form data", From 896ee39f68299d2184592290a7c6d9d8f4c12ee9 Mon Sep 17 00:00:00 2001 From: marius-mather <61438033+marius-mather@users.noreply.github.com> Date: Wed, 29 Jul 2026 14:37:25 +1000 Subject: [PATCH 62/63] feat: track service unit cost of runs in database (SBP-445) (#113) * feat: add service_usage column to WorkflowRun * feat: function to sync service usage for a run * fix: check current service usage is not None * refactor: rework workflow results so we can get service usage * feat: sync service usage when looking up jobs (current syncs trigger here) * test: update tests with reworked results/outputs * chore: add httpx2, recommended by FastAPI/starlette * test: unit test of service usage calculation * test: add test factory for S3Object * test: add test of extracting service units for run * fix: update arg name in rfdiffusion spec * fix: linting * test: unit tests of read_s3_bytes * test: more tests of result_utils to improve coverage * chore: linting * chore: add service_usage to workflowrun admin --- ...60728_145217_service_usage_4f26ab5038ec.py | 23 ++ app/db/admin.py | 3 +- app/db/models/core.py | 2 + app/routes/workflow/jobs.py | 8 + app/services/job_utils.py | 24 ++ app/services/results_utils.py | 96 ++++++- pyproject.toml | 1 + tests/conftest.py | 1 + tests/datagen.py | 6 +- tests/test_routes_workflow_jobs_extra.py | 4 +- tests/test_services_results_utils.py | 243 +++++++++++++++++- tests/test_services_s3.py | 38 +++ uv.lock | 39 +++ 13 files changed, 464 insertions(+), 24 deletions(-) create mode 100644 alembic/versions/20260728_145217_service_usage_4f26ab5038ec.py diff --git a/alembic/versions/20260728_145217_service_usage_4f26ab5038ec.py b/alembic/versions/20260728_145217_service_usage_4f26ab5038ec.py new file mode 100644 index 0000000..8935f18 --- /dev/null +++ b/alembic/versions/20260728_145217_service_usage_4f26ab5038ec.py @@ -0,0 +1,23 @@ +"""service_usage""" + +from alembic import op +import sqlalchemy as sa + + +# revision identifiers, used by Alembic. +revision = '4f26ab5038ec' +down_revision = 'a6ee1af2c156' +branch_labels = None +depends_on = None + + +def upgrade() -> None: + # ### commands auto generated by Alembic - please adjust! ### + op.add_column('workflow_runs', sa.Column('service_usage', sa.Float(), nullable=True)) + # ### end Alembic commands ### + + +def downgrade() -> None: + # ### commands auto generated by Alembic - please adjust! ### + op.drop_column('workflow_runs', 'service_usage') + # ### end Alembic commands ### diff --git a/app/db/admin.py b/app/db/admin.py index 04fad8a..6326298 100644 --- a/app/db/admin.py +++ b/app/db/admin.py @@ -168,8 +168,9 @@ class WorkflowRunAdmin(ModelView): HasOne("owner", identity="app-user"), "seqera_run_id", "run_name", - "binder_name", + "service_usage", JSONField("submitted_form_data"), + "binder_name", "work_dir", "submission_timestamp", ] diff --git a/app/db/models/core.py b/app/db/models/core.py index 8440057..ac5736f 100644 --- a/app/db/models/core.py +++ b/app/db/models/core.py @@ -7,6 +7,7 @@ JSON, BigInteger, DateTime, + Float, ForeignKey, Index, Numeric, @@ -92,6 +93,7 @@ class WorkflowRun(Base): DateTime(timezone=True), nullable=True ) tool: Mapped[str | None] = mapped_column(Text, nullable=True) + service_usage: Mapped[float | None] = mapped_column(Float, nullable=True) owner: Mapped[AppUser] = relationship(back_populates="workflow_runs") workflow: Mapped[Workflow | None] = relationship(back_populates="runs") diff --git a/app/routes/workflow/jobs.py b/app/routes/workflow/jobs.py index 8824fd6..40bf523 100644 --- a/app/routes/workflow/jobs.py +++ b/app/routes/workflow/jobs.py @@ -33,6 +33,7 @@ get_owned_run_by_id, get_user_job_list_rows, parse_submit_datetime, + sync_service_usage, ) from ...services.seqera import describe_workflow from ...services.seqera_client import cancel_workflow_raw, delete_workflow_raw, delete_workflows_raw @@ -200,6 +201,10 @@ async def _fetch_seqera(user_run: UserJobListRow) -> tuple[str, dict[str, object if score is None: score = await ensure_completed_run_score(db, owned_run, ui_status) + # Also sync service usage for runs, if needed + if owned_run.service_usage is None: + await sync_service_usage(db, run=owned_run, ui_status=ui_status) + jobs.append( JobListItem( id=run_id, @@ -259,6 +264,9 @@ async def get_job_details( score = await ensure_completed_run_score(db, owned_run, ui_status) if ui_status != "Completed": score = None + # Sync service usage - not returned to user but currently this is where + # we trigger syncs + await sync_service_usage(db, run=owned_run, ui_status=ui_status) raw_tool: str | None = getattr(owned_run, "tool", None) or None if not raw_tool: diff --git a/app/services/job_utils.py b/app/services/job_utils.py index 3015d95..c01e775 100644 --- a/app/services/job_utils.py +++ b/app/services/job_utils.py @@ -206,3 +206,27 @@ async def ensure_completed_run_score(db: Session, run: WorkflowRun, ui_status: s db.add(RunMetric(run_id=run.id, max_score=bounded_score)) db.commit() return _round_score(bounded_score) + + +async def sync_service_usage(db: Session, run: WorkflowRun, ui_status: str) -> float | None: + if ui_status != "Completed": + return None + if run.service_usage is not None: + return run.service_usage + + try: + spec = get_output_spec(run) + except ValueError as exc: + logger.warning(f"Can't get service usage for run {run.id} - can't get output spec: {exc}") + return None + + await sync_workflow_outputs(db, run=run, spec=spec, suppress_s3_errors=True) + + usage = await spec.get_service_units(db, run) + if usage is None: + return None + else: + run.service_usage = usage + db.add(run) + db.commit() + return usage diff --git a/app/services/results_utils.py b/app/services/results_utils.py index 42eb2f7..0f98c7b 100644 --- a/app/services/results_utils.py +++ b/app/services/results_utils.py @@ -27,7 +27,7 @@ read_s3_file, ) -OutputCategory = Literal["report", "stats_csv", "pdb", "snapshot", "alignment"] +OutputCategory = Literal["report", "stats_csv", "pdb", "snapshot", "alignment", "usage"] class OutputClassifier(Protocol): @@ -59,11 +59,22 @@ class WorkflowResultsSpec: tool: WorkflowTool required_categories: set[OutputCategory] get_prefixes: Callable[[WorkflowRun], list[str]] - classify: OutputClassifier + classifier: OutputClassifier get_score_file: GetScoreFile extract_max_score: Callable[[str], Awaitable[float | None]] supports_snapshots: bool = False + def classify_output(self, key: str, sample_id: str | None) -> ClassifiedOutput | None: + """ + Classify a workflow output into a category. + First checks for outputs that are shared across workflows, + then the unique outputs for the specific workflow. + """ + shared_output = classify_shared_outputs(key) + if shared_output is not None: + return shared_output + return self.classifier(key, sample_id) + async def get_max_score(self, db: Session, run: WorkflowRun): keys = _get_run_output_keys(db, run) sample_id = get_sample_id_for_result(run) @@ -76,6 +87,27 @@ async def get_max_score(self, db: Session, run: WorkflowRun): logger.warning("Failed to extract max score from %r: %s", score_file, e, exc_info=True) return None + def get_usage_file(self, keys: list[str]) -> str | None: + for key in keys: + normalized = key.strip() + if not normalized: + continue + basename = normalized.rsplit("/", 1)[-1] + if basename == "UsageReport.csv": + return normalized + return None + + async def get_service_units(self, db: Session, run: WorkflowRun) -> float | None: + keys = _get_run_output_keys(db, run) + usage_file = self.get_usage_file(keys) + if usage_file is None: + return None + try: + return await get_run_service_usage(usage_file) + except Exception as e: + logger.warning("Failed to get service usage from %r: %s", usage_file, e, exc_info=True) + return None + _LOG_LEVEL_PATTERN = re.compile(r"\b(TRACE|DEBUG|INFO|WARN|WARNING|ERROR|FATAL)\b") _LOG_TIMESTAMP_PATTERN = re.compile( @@ -88,6 +120,22 @@ async def get_max_score(self, db: Session, run: WorkflowRun): logger = logging.getLogger(__name__) +def classify_shared_outputs(key: str) -> ClassifiedOutput | None: + """ + Classify outputs that are shared across workflows. Currently + only the usage report, but other common outputs can be added + """ + normalized = key.strip() + if not normalized or normalized.endswith("/"): + return None + + basename = normalized.rsplit("/", 1)[-1] + if basename == "UsageReport.csv": + return ClassifiedOutput(category="usage", label=basename) + + return None + + def _sanitize_content_disposition_filename(filename: str) -> str: sanitized = _HEADER_UNSAFE_FILENAME_CHARS.sub("_", filename).strip() return sanitized or "download" @@ -511,6 +559,7 @@ def build_bindcraft_output_listing_prefixes(run: WorkflowRun) -> list[str]: # Always include run-UUID-only prefixes; these do not depend on sample_id. prefixes: list[str] = [ + f"{run_uuid}/", f"{run_uuid}/ranker/", f"{run_uuid}/generate/", ] @@ -535,6 +584,7 @@ def build_boltz_proteinfold_output_listing_prefixes(run: WorkflowRun) -> list[st sample_id = get_sample_id_for_result(run) prefixes = [ + f"{run_uuid}/", f"{run_uuid}/reports/", f"{run_uuid}/boltz/top_ranked_structures/", f"{run_uuid}/mmseqs/", @@ -556,6 +606,7 @@ def build_alphafold2_proteinfold_output_listing_prefixes(run: WorkflowRun) -> li sample_id = get_sample_id_for_result(run) prefixes = [ + f"{run_uuid}/", f"{run_uuid}/reports/", f"{run_uuid}/alphafold2/split_msa_prediction/top_ranked_structures/", ] @@ -627,6 +678,7 @@ def build_wisps_output_listing_prefixes(run: WorkflowRun) -> list[str]: if not run_uuid: return [] return [ + f"{run_uuid}/", f"{run_uuid}/multiqc/", f"{run_uuid}/collect/", f"{run_uuid}/ipsae/", @@ -641,6 +693,7 @@ def build_colabfold_proteinfold_output_listing_prefixes(run: WorkflowRun) -> lis sample_id = get_sample_id_for_result(run) prefixes = [ + f"{run_uuid}/", f"{run_uuid}/reports/", f"{run_uuid}/colabfold/top_ranked_structures/", f"{run_uuid}/mmseqs/", @@ -697,7 +750,7 @@ def build_rfdiffusion_output_listing_prefixes(run: WorkflowRun) -> list[str]: run_uuid = str(getattr(run, "id", "") or "").strip() if not run_uuid: return [] - return [f"{run_uuid}/results/"] + return [f"{run_uuid}/", f"{run_uuid}/results/"] def _make_wisps_spec(tool: WorkflowTool) -> WorkflowResultsSpec: @@ -708,7 +761,7 @@ def _make_wisps_spec(tool: WorkflowTool) -> WorkflowResultsSpec: get_prefixes=build_wisps_output_listing_prefixes, get_score_file=get_wisps_score_file, extract_max_score=extract_wisps_max_score, - classify=classify_wisps_output_key, + classifier=classify_wisps_output_key, ) @@ -720,7 +773,7 @@ def _make_bulk_prediction_spec(tool: WorkflowTool) -> WorkflowResultsSpec: get_prefixes=build_wisps_output_listing_prefixes, get_score_file=get_wisps_score_file, extract_max_score=extract_bulk_prediction_max_score, - classify=classify_wisps_output_key, + classifier=classify_wisps_output_key, ) @@ -733,7 +786,7 @@ def _make_bulk_prediction_spec(tool: WorkflowTool) -> WorkflowResultsSpec: get_prefixes=build_bindcraft_output_listing_prefixes, get_score_file=get_bindcraft_score_file, extract_max_score=extract_bindcraft_max_score, - classify=classify_bindcraft_output_key, + classifier=classify_bindcraft_output_key, supports_snapshots=True, ), "rfdiffusion": WorkflowResultsSpec( @@ -743,7 +796,7 @@ def _make_bulk_prediction_spec(tool: WorkflowTool) -> WorkflowResultsSpec: get_prefixes=build_rfdiffusion_output_listing_prefixes, get_score_file=get_rfdiffusion_score_file, extract_max_score=extract_rfdiffusion_max_score, - classify=classify_rfdiffusion_output_key, + classifier=classify_rfdiffusion_output_key, ), }, "single-prediction": { @@ -754,7 +807,7 @@ def _make_bulk_prediction_spec(tool: WorkflowTool) -> WorkflowResultsSpec: get_prefixes=build_boltz_proteinfold_output_listing_prefixes, get_score_file=get_proteinfold_score_file, extract_max_score=extract_proteinfold_max_score, - classify=classify_boltz_proteinfold_output, + classifier=classify_boltz_proteinfold_output, ), "alphafold2": WorkflowResultsSpec( kind="single-prediction", @@ -763,7 +816,7 @@ def _make_bulk_prediction_spec(tool: WorkflowTool) -> WorkflowResultsSpec: get_prefixes=build_alphafold2_proteinfold_output_listing_prefixes, get_score_file=get_proteinfold_score_file, extract_max_score=extract_proteinfold_max_score, - classify=classify_alphafold2_proteinfold_output, + classifier=classify_alphafold2_proteinfold_output, ), "colabfold": WorkflowResultsSpec( kind="single-prediction", @@ -772,7 +825,7 @@ def _make_bulk_prediction_spec(tool: WorkflowTool) -> WorkflowResultsSpec: get_prefixes=build_colabfold_proteinfold_output_listing_prefixes, get_score_file=get_proteinfold_score_file, extract_max_score=extract_proteinfold_max_score, - classify=classify_colabfold_proteinfold_output, + classifier=classify_colabfold_proteinfold_output, ), }, "interaction-screening": { @@ -802,7 +855,7 @@ def collect_classified_outputs( outputs = {} sample_id = get_sample_id_for_result(run) for key in _get_run_output_keys(db, run): - classified = spec.classify(key, sample_id) + classified = spec.classify_output(key, sample_id) if classified: outputs[key] = classified return outputs @@ -881,7 +934,7 @@ async def list_workflow_outputs_from_s3( if not key or key in outputs: continue - classified = spec.classify(key, sample_id) + classified = spec.classify_output(key, sample_id) if classified is not None: outputs[key] = classified @@ -966,7 +1019,7 @@ async def get_result_output_downloads(db: Session, run: WorkflowRun) -> list[Res # Sort and filter outputs for key, output in sorted(outputs.items(), key=_get_output_sort_key): - if output.category == "snapshot": + if output.category in ("snapshot", "usage"): continue downloads.append( ResultDownloadItem( @@ -990,7 +1043,11 @@ async def get_all_downloads_zipped(db: Session, run: WorkflowRun) -> BytesIO: results_spec = get_output_spec(run) outputs = collect_classified_outputs(db, run, results_spec) # Exclude snapshots - downloads = [(key, output) for key, output in outputs.items() if output.category != "snapshot"] + downloads = [ + (key, output) + for key, output in outputs.items() + if output.category not in ("snapshot", "usage") + ] used_filenames: set[str] = set() zip_file = BytesIO() @@ -1083,3 +1140,14 @@ async def get_result_snapshot_downloads(db: Session, run: WorkflowRun) -> list[R ) ) return downloads + + +async def get_run_service_usage(usage_file: str) -> float | None: + content = await read_s3_file(usage_file) + csv_reader = csv.DictReader(StringIO(content)) + su_values: list[float] = [] + for row in csv_reader: + value = row.get("Service Units") + if value: + su_values.append(float(value.strip())) + return sum(su_values) diff --git a/pyproject.toml b/pyproject.toml index 91b0591..67a813c 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -41,6 +41,7 @@ dev = [ "pytest-cov~=4.1", "pytest-mock~=3.12", "httpx~=0.28", + "httpx2~=2.9", "coverage[toml]~=7.4", "ruff~=0.14", "black~=26.5", diff --git a/tests/conftest.py b/tests/conftest.py index e1f0a0b..6f99103 100644 --- a/tests/conftest.py +++ b/tests/conftest.py @@ -140,6 +140,7 @@ def persistent_models(test_db): datagen.WorkflowRunFactory, datagen.RunInputFactory, datagen.RunOutputFactory, + datagen.S3ObjectFactory, datagen.QueuedJobFactory, ] diff --git a/tests/datagen.py b/tests/datagen.py index 11fd9af..e9db6e4 100644 --- a/tests/datagen.py +++ b/tests/datagen.py @@ -5,7 +5,7 @@ from polyfactory.factories.dataclass_factory import DataclassFactory from polyfactory.factories.sqlalchemy_factory import SQLAlchemyFactory -from app.db.models.core import AppUser, RunInput, RunOutput, Workflow, WorkflowRun +from app.db.models.core import AppUser, RunInput, RunOutput, S3Object, Workflow, WorkflowRun from app.db.models.job_queue import QueuedJob from app.services.job_utils import UserJobListRow @@ -62,6 +62,10 @@ class RunOutputFactory(SQLAlchemyFactory[RunOutput]): __set_relationships__ = False +class S3ObjectFactory(SQLAlchemyFactory[S3Object]): + __set_relationships__ = False + + class QueuedJobFactory(SQLAlchemyFactory[QueuedJob]): __set_relationships__ = False diff --git a/tests/test_routes_workflow_jobs_extra.py b/tests/test_routes_workflow_jobs_extra.py index 959ae81..4ed1537 100644 --- a/tests/test_routes_workflow_jobs_extra.py +++ b/tests/test_routes_workflow_jobs_extra.py @@ -96,7 +96,7 @@ async def test_get_job_details_success(test_db): ) workflow = Workflow( id=uuid4(), - name="BindCraft", + name="de-novo-design", description="Binding workflow", repo_url="https://github.com/test/bindcraft", default_revision="main", @@ -136,7 +136,7 @@ async def test_get_job_details_success(test_db): assert result.id == str(run.id) assert result.jobName == "PDL1" - assert result.workflow == "Bindcraft" + assert result.workflow == "De Novo Design" assert result.score == 0.912 diff --git a/tests/test_services_results_utils.py b/tests/test_services_results_utils.py index 714054d..9db273d 100644 --- a/tests/test_services_results_utils.py +++ b/tests/test_services_results_utils.py @@ -4,7 +4,7 @@ from io import BytesIO from types import SimpleNamespace -from unittest.mock import AsyncMock, patch +from unittest.mock import AsyncMock, MagicMock, patch from uuid import uuid4 from zipfile import ZipFile @@ -20,29 +20,36 @@ build_bindcraft_output_listing_prefixes, build_boltz_proteinfold_output_listing_prefixes, build_colabfold_proteinfold_output_listing_prefixes, + build_rfdiffusion_output_listing_prefixes, build_wisps_output_listing_prefixes, classify_alphafold2_proteinfold_output, classify_bindcraft_output_key, classify_boltz_proteinfold_output, classify_colabfold_proteinfold_output, + classify_rfdiffusion_output_key, + classify_shared_outputs, classify_wisps_output_key, extract_bindcraft_max_score, extract_proteinfold_max_score, + extract_rfdiffusion_max_score, extract_wisps_max_score, format_log_entries, get_all_downloads_zipped, get_bindcraft_score_file, get_proteinfold_score_file, + get_rfdiffusion_score_file, + get_run_service_usage, get_sample_id_for_result, get_tool_name, get_wisps_score_file, + list_workflow_outputs_from_s3, resolve_fasta_form_data, resolve_pdb_presigned_urls, resolve_submitted_form_data, s3_uri_to_key, ) from app.services.s3 import S3ServiceError -from tests.datagen import AppUserFactory, WorkflowRunFactory +from tests.datagen import AppUserFactory, RunOutputFactory, S3ObjectFactory, WorkflowRunFactory def test_format_log_entries_extracts_timestamp_level_and_strips_ansi(): @@ -210,6 +217,7 @@ def test_bindcraft_helpers_classify_keys_and_build_prefixes(monkeypatch): prefixes = build_bindcraft_output_listing_prefixes(run) assert prefixes == [ + f"{run.id}/", f"{run.id}/ranker/", f"{run.id}/generate/", f"{run.id}/bindcraft/sampleZ_0_output/", @@ -217,6 +225,7 @@ def test_bindcraft_helpers_classify_keys_and_build_prefixes(monkeypatch): run_without_sample = SimpleNamespace(id=run.id, sample_id=None, binder_name=None, form_id=None) assert build_bindcraft_output_listing_prefixes(run_without_sample) == [ + f"{run.id}/", f"{run.id}/ranker/", f"{run.id}/generate/", ] @@ -227,6 +236,81 @@ def test_bindcraft_helpers_classify_keys_and_build_prefixes(monkeypatch): assert _build_s3_uri("path/to/file.txt") == "path/to/file.txt" +def test_rfdiffusion_helpers_classify_keys_and_build_prefixes(): + run = WorkflowRun(id=uuid4(), owner_user_id=uuid4(), sample_id="sampleZ") + + assert classify_rfdiffusion_output_key(" ") is None + assert classify_rfdiffusion_output_key("folder/") is None + assert classify_rfdiffusion_output_key(f"{run.id}/ranked_designs.csv") is None + assert classify_rfdiffusion_output_key(f"{run.id}/results/other.txt") is None + assert classify_rfdiffusion_output_key( + f"{run.id}/results/ranked_designs.csv" + ) == ClassifiedOutput( + "stats_csv", + "ranked_designs.csv", + ) + assert classify_rfdiffusion_output_key( + f"{run.id}/results/ranked_designs.tar.gz" + ) == ClassifiedOutput( + "pdb", + "ranked_designs.tar.gz", + ) + + assert build_rfdiffusion_output_listing_prefixes(run) == [ + f"{run.id}/", + f"{run.id}/results/", + ] + assert build_rfdiffusion_output_listing_prefixes(SimpleNamespace(id=None)) == [] + + +def test_get_rfdiffusion_score_file_uses_ranked_designs_csv(): + assert ( + get_rfdiffusion_score_file( + [ + " ", + "run-1/results/other.csv", + " run-1/results/ranked_designs.csv ", + ], + sample_id="sampleZ", + ) + == "run-1/results/ranked_designs.csv" + ) + assert get_rfdiffusion_score_file(["run-1/results/other.csv"], sample_id=None) is None + + +@pytest.mark.asyncio +async def test_extract_rfdiffusion_max_score_reads_first_ranked_design_score(): + csv_text = "design,af2_plddt_overall\nsampleZ_0,91.3\nsampleZ_1,88.1\n" + + with patch( + "app.services.results_utils.read_s3_file", + new_callable=AsyncMock, + return_value=csv_text, + ) as read_file: + score = await extract_rfdiffusion_max_score("run-1/results/ranked_designs.csv") + + assert score == pytest.approx(0.913) + read_file.assert_awaited_once_with("run-1/results/ranked_designs.csv") + + +@pytest.mark.asyncio +async def test_extract_rfdiffusion_max_score_returns_none_without_score_value(): + """ + Test extract_rfdiffusion_max_score returns None when no af2_plddt_overall score is available + """ + for csv_text in [ + "design,af2_plddt_overall\n", + "design,af2_plddt_overall\nsampleZ_0, \n", + "design,other_score\nsampleZ_0,91.3\n", + ]: + with patch( + "app.services.results_utils.read_s3_file", + new_callable=AsyncMock, + return_value=csv_text, + ): + assert await extract_rfdiffusion_max_score("run-1/results/ranked_designs.csv") is None + + @pytest.mark.asyncio async def test_get_all_downloads_zipped_writes_category_label_files_and_reads_each_output( test_db, persistent_models @@ -352,6 +436,51 @@ def test_all_workflow_output_specs_have_score_hooks(): assert spec.extract_max_score is not None +@pytest.mark.asyncio +async def test_list_workflow_outputs_from_s3_continues(): + """ + Test list_wokflow_outputs_from_s3 continues past S3 errors when suppress_s3_errors is True + """ + run = WorkflowRun( + id=uuid4(), + owner_user_id=uuid4(), + sample_id="sampleZ", + seqera_run_id="seqera-run-1", + ) + + async def extract_max_score(_key: str) -> float | None: + return None + + spec = WorkflowResultsSpec( + kind="de-novo-design", + tool="rfdiffusion", + required_categories=set(), + get_prefixes=lambda _run: ["failed-prefix/", "ok-prefix/"], + classifier=classify_rfdiffusion_output_key, + get_score_file=lambda _keys, _sample_id: None, + extract_max_score=extract_max_score, + ) + + with patch( + "app.services.results_utils.list_s3_files", + new=AsyncMock( + side_effect=[ + S3ServiceError("failed to list"), + [{"key": "ok-prefix/results/ranked_designs.csv"}], + ] + ), + ) as list_s3_files_mock: + outputs = await list_workflow_outputs_from_s3(run, spec, suppress_s3_errors=True) + + assert outputs == { + "ok-prefix/results/ranked_designs.csv": ClassifiedOutput( + category="stats_csv", + label="ranked_designs.csv", + ) + } + assert list_s3_files_mock.await_count == 2 + + @pytest.mark.asyncio async def test_workflow_results_spec_get_max_score_returns_none_without_score_file(test_db): user = AppUser( @@ -374,7 +503,7 @@ async def test_workflow_results_spec_get_max_score_returns_none_without_score_fi tool="boltz", required_categories=set(), get_prefixes=lambda _run: [], - classify=lambda _key, _sample_id: None, + classifier=lambda _key, _sample_id: None, get_score_file=lambda _keys, _sample_id: None, extract_max_score=extractor, ) @@ -420,7 +549,7 @@ async def test_workflow_results_spec_get_max_score_extracts_selected_run_output( tool="boltz", required_categories=set(), get_prefixes=lambda _run: [], - classify=lambda _key, _sample_id: None, + classifier=lambda _key, _sample_id: None, get_score_file=get_proteinfold_score_file, extract_max_score=extractor, ) @@ -429,6 +558,50 @@ async def test_workflow_results_spec_get_max_score_extracts_selected_run_output( extractor.assert_awaited_once_with("run-1/boltz/T1024/T1024_ptm.tsv") +@pytest.mark.asyncio +async def test_workflow_results_spec_get_service_units_uses_usage_report_key( + test_db, persistent_models +): + user = AppUserFactory.create_sync() + run = WorkflowRunFactory.create_sync( + owner=user, + work_dir="workdir-usage-selected", + ) + ignored = S3ObjectFactory.create_sync( + object_key="run-1/boltz/T1024/T1024_ptm.tsv", + uri="s3://bucket/run-1/boltz/T1024/T1024_ptm.tsv", + ) + usage_object = S3ObjectFactory.create_sync( + object_key="run-1/UsageReport.csv", + uri="s3://bucket/run-1/UsageReport.csv", + ) + RunOutputFactory.create_sync(run_id=run.id, s3_object_id=ignored.object_key) + RunOutputFactory.create_sync( + run_id=run.id, + s3_object_id=usage_object.object_key, + ) + + spec = WorkflowResultsSpec( + kind="single-prediction", + tool="boltz", + required_categories=set(), + get_prefixes=lambda _run: [], + classifier=lambda _key, _sample_id: None, + get_score_file=get_proteinfold_score_file, + extract_max_score=AsyncMock(return_value=0.91), + ) + + with patch( + "app.services.results_utils.get_run_service_usage", + new_callable=AsyncMock, + return_value=3.45, + ) as get_run_service_usage_mock: + service_units = await spec.get_service_units(test_db, run) + + assert service_units == 3.45 + get_run_service_usage_mock.assert_awaited_once_with("run-1/UsageReport.csv") + + def test_boltz_proteinfold_helpers_classify_keys_and_build_prefixes(): run = WorkflowRun(id=uuid4(), owner_user_id=uuid4(), sample_id="T1024") @@ -482,6 +655,7 @@ def test_boltz_proteinfold_helpers_classify_keys_and_build_prefixes(): ) == ClassifiedOutput("alignment", "single_prediction.a3m") assert build_boltz_proteinfold_output_listing_prefixes(run) == [ + f"{run.id}/", f"{run.id}/reports/", f"{run.id}/boltz/top_ranked_structures/", f"{run.id}/mmseqs/", @@ -538,6 +712,7 @@ def test_alphafold2_proteinfold_helpers_classify_keys_and_build_prefixes(): ) == ClassifiedOutput("stats_csv", "T1024_0_pae.tsv") assert build_alphafold2_proteinfold_output_listing_prefixes(run) == [ + f"{run.id}/", f"{run.id}/reports/", f"{run.id}/alphafold2/split_msa_prediction/top_ranked_structures/", f"{run.id}/alphafold2/split_msa_prediction/T1024/", @@ -597,6 +772,7 @@ def test_colabfold_proteinfold_helpers_classify_keys_and_build_prefixes(): ) == ClassifiedOutput("alignment", "single_prediction.a3m") assert build_colabfold_proteinfold_output_listing_prefixes(run) == [ + f"{run.id}/", f"{run.id}/reports/", f"{run.id}/colabfold/top_ranked_structures/", f"{run.id}/mmseqs/", @@ -889,15 +1065,41 @@ async def test_extract_wisps_max_score_skips_non_numeric_iptm(): assert score == 0.75 +# --------------------------------------------------------------------------- +# get_run_service_usage +# --------------------------------------------------------------------------- + + +@pytest.mark.asyncio +async def test_get_run_service_usage(): + csv_text = ( + "Name,Queue,Service Units\n" + "prepare,normalbw,0.00\n" + "search,normalbw,0.20\n" + "predict,gpuhopper,3.25\n" + ) + + with patch( + "app.services.results_utils.read_s3_file", + new_callable=AsyncMock, + return_value=csv_text, + ) as mock_read: + service_usage = await get_run_service_usage("run-1/UsageReport.csv") + + assert service_usage == 3.45 + mock_read.assert_awaited_once_with("run-1/UsageReport.csv") + + # --------------------------------------------------------------------------- # build_wisps_output_listing_prefixes # --------------------------------------------------------------------------- -def test_build_wisps_output_listing_prefixes_returns_three_prefixes(): +def test_build_wisps_output_listing_prefixes(): run = WorkflowRun(id=uuid4(), owner_user_id=uuid4(), sample_id="sample1") prefixes = build_wisps_output_listing_prefixes(run) - assert len(prefixes) == 3 + assert len(prefixes) == 4 + assert f"{run.id}/" in prefixes assert f"{run.id}/multiqc/" in prefixes assert f"{run.id}/collect/" in prefixes assert f"{run.id}/ipsae/" in prefixes @@ -922,3 +1124,32 @@ def test_workflow_output_specs_has_interaction_screening(): assert spec.kind == "interaction-screening" assert spec.get_score_file is not None assert spec.extract_max_score is not None + + +def test_classify_shared_outputs_usage_report_csv(): + result = classify_shared_outputs("runs/run-1/UsageReport.csv") + + assert result == ClassifiedOutput(category="usage", label="UsageReport.csv") + + +def test_workflow_results_spec_classify_output_uses_shared_outputs(): + """Test WorkflowResultsSpec classifies shared outputs before workflow-specific outputs.""" + + async def extract_max_score(_key: str) -> float | None: + return None + + classifier = MagicMock(return_value=ClassifiedOutput(category="pdb", label="ignored.pdb")) + spec = WorkflowResultsSpec( + kind="single-prediction", + tool="boltz", + required_categories=set(), + get_prefixes=lambda _run: [], + classifier=classifier, + get_score_file=lambda _keys, _sample_id: None, + extract_max_score=extract_max_score, + ) + + result = spec.classify_output("runs/run-1/UsageReport.csv", sample_id="T1024") + + assert result == ClassifiedOutput(category="usage", label="UsageReport.csv") + classifier.assert_not_called() diff --git a/tests/test_services_s3.py b/tests/test_services_s3.py index 8fa5b41..821ddce 100644 --- a/tests/test_services_s3.py +++ b/tests/test_services_s3.py @@ -17,6 +17,7 @@ get_s3_client, list_s3_files, read_csv_from_s3, + read_s3_bytes, read_s3_file, upload_file_to_s3, ) @@ -265,6 +266,43 @@ async def test_read_s3_file_client_error(mock_env_vars, mock_s3_client): await read_s3_file("results/test/missing.tsv") +@pytest.mark.asyncio +async def test_read_s3_bytes_success(mock_env_vars, mock_s3_client): + """Test reading a bytes object from S3.""" + content = b"rank\tscore\n0\t0.91\n" + mock_response = {"Body": MagicMock()} + mock_response["Body"].read.return_value = content + mock_s3_client.get_object.return_value = mock_response + + result = await read_s3_bytes("results/test/file.tsv") + + assert result == content + mock_s3_client.get_object.assert_called_once_with( + Bucket="test-bucket", + Key="results/test/file.tsv", + ) + + +@pytest.mark.asyncio +async def test_read_s3_bytes_missing_bucket(): + """Test reading a text object without bucket configuration.""" + with patch.dict("os.environ", {}, clear=True): + with pytest.raises(S3ConfigurationError, match="AWS_S3_BUCKET"): + await read_s3_bytes("results/test/file.tsv") + + +@pytest.mark.asyncio +async def test_read_s3_bytes_client_error(mock_env_vars, mock_s3_client): + """Test reading a text object wraps S3 client errors.""" + mock_s3_client.get_object.side_effect = ClientError( + {"Error": {"Code": "NoSuchKey", "Message": "Not found"}}, + "get_object", + ) + + with pytest.raises(S3ServiceError, match="Failed to read file from S3"): + await read_s3_bytes("results/test/missing.tsv") + + @pytest.mark.asyncio async def test_read_csv_from_s3_all_columns(mock_env_vars, mock_s3_client): """Test reading CSV with all columns.""" diff --git a/uv.lock b/uv.lock index ec6b5d1..1d6be17 100644 --- a/uv.lock +++ b/uv.lock @@ -397,6 +397,19 @@ wheels = [ { url = "https://files.pythonhosted.org/packages/7e/f5/f66802a942d491edb555dd61e3a9961140fd64c90bce1eafd741609d334d/httpcore-1.0.9-py3-none-any.whl", hash = "sha256:2d400746a40668fc9dec9810239072b40b4484b640a8c38fd654a024c7a1bf55", size = 78784, upload-time = "2025-04-24T22:06:20.566Z" }, ] +[[package]] +name = "httpcore2" +version = "2.9.1" +source = { registry = "https://pypi.org/simple" } +dependencies = [ + { name = "h11" }, + { name = "truststore" }, +] +sdist = { url = "https://files.pythonhosted.org/packages/39/a8/20ed1ed79cbc2ecdf5301c0968ab7c85547212e2a7bd126ddd2d986e206e/httpcore2-2.9.1.tar.gz", hash = "sha256:4d8acbf8b306f48c9d6046591fd5ba4037d1b1b1000d140fc2c3eab1e9a0c0e2", size = 67089, upload-time = "2026-07-24T09:21:03.867Z" } +wheels = [ + { url = "https://files.pythonhosted.org/packages/9f/fb/46c52b781975c335a2bcf1072c7bbc007cbdc8d674217f5ee1daba2c848b/httpcore2-2.9.1-py3-none-any.whl", hash = "sha256:6182472379e855fe4221246a2bb7ecede403bc61c6798062ae1787d051ccde26", size = 82809, upload-time = "2026-07-24T09:21:01.178Z" }, +] + [[package]] name = "httptools" version = "0.8.0" @@ -434,6 +447,21 @@ wheels = [ { url = "https://files.pythonhosted.org/packages/2a/39/e50c7c3a983047577ee07d2a9e53faf5a69493943ec3f6a384bdc792deb2/httpx-0.28.1-py3-none-any.whl", hash = "sha256:d909fcccc110f8c7faf814ca82a9a4d816bc5a6dbfea25d6591d6985b8ba59ad", size = 73517, upload-time = "2024-12-06T15:37:21.509Z" }, ] +[[package]] +name = "httpx2" +version = "2.9.1" +source = { registry = "https://pypi.org/simple" } +dependencies = [ + { name = "anyio" }, + { name = "httpcore2" }, + { name = "idna" }, + { name = "truststore" }, +] +sdist = { url = "https://files.pythonhosted.org/packages/21/14/38128fbafd7e0ed41d874df6c9a653d47c2d111cfe59e2b4ac95161b4abd/httpx2-2.9.1.tar.gz", hash = "sha256:1932a768737e3666291582833da748cc4e563c337cf96706fccc04fa6e58764a", size = 95458, upload-time = "2026-07-24T09:21:04.972Z" } +wheels = [ + { url = "https://files.pythonhosted.org/packages/13/b8/cfd91c4ab9134d386d48f0b6ac662ff3d4be6efdee59ee1c67ebc3c0487c/httpx2-2.9.1-py3-none-any.whl", hash = "sha256:1820fe14a9ab1107bfeff39259987429450b070ec0ff38cc87eb0d8c97fdc71a", size = 91191, upload-time = "2026-07-24T09:21:02.6Z" }, +] + [[package]] name = "idna" version = "3.18" @@ -1095,6 +1123,7 @@ dev = [ { name = "faker" }, { name = "groovy-parser" }, { name = "httpx" }, + { name = "httpx2" }, { name = "mypy" }, { name = "polyfactory" }, { name = "pytest" }, @@ -1137,6 +1166,7 @@ dev = [ { name = "faker", specifier = ">=40.1.0" }, { name = "groovy-parser", git = "https://github.com/inab/python-groovy-parser.git?rev=5e7f3a4e7dcc5fb2e186fe7ca30df260ce5efeb8" }, { name = "httpx", specifier = "~=0.28" }, + { name = "httpx2", specifier = "~=2.9" }, { name = "mypy", specifier = "~=1.19" }, { name = "polyfactory", specifier = "~=2.16" }, { name = "pytest", specifier = "~=8.3" }, @@ -1234,6 +1264,15 @@ wheels = [ { url = "https://files.pythonhosted.org/packages/5d/7c/add7d4be6690fb61f4d30b672e544b1ac541c824f05a7de50ff5d8985db5/starlette_admin-0.16.1-py3-none-any.whl", hash = "sha256:a5e6cb0beb2e9bdc65b07dbb4bd01726aaad34764595e3bc3ff8578e698ff98a", size = 2176514, upload-time = "2026-06-06T20:10:42.308Z" }, ] +[[package]] +name = "truststore" +version = "0.10.4" +source = { registry = "https://pypi.org/simple" } +sdist = { url = "https://files.pythonhosted.org/packages/53/a3/1585216310e344e8102c22482f6060c7a6ea0322b63e026372e6dcefcfd6/truststore-0.10.4.tar.gz", hash = "sha256:9d91bd436463ad5e4ee4aba766628dd6cd7010cf3e2461756b3303710eebc301", size = 26169, upload-time = "2025-08-12T18:49:02.73Z" } +wheels = [ + { url = "https://files.pythonhosted.org/packages/19/97/56608b2249fe206a67cd573bc93cd9896e1efb9e98bce9c163bcdc704b88/truststore-0.10.4-py3-none-any.whl", hash = "sha256:adaeaecf1cbb5f4de3b1959b42d41f6fab57b2b1666adb59e89cb0b53361d981", size = 18660, upload-time = "2025-08-12T18:49:01.46Z" }, +] + [[package]] name = "typer" version = "0.26.8" From 4f03a2bcacebb49cade3407f37a8e6c5a127538d Mon Sep 17 00:00:00 2001 From: marius-mather <61438033+marius-mather@users.noreply.github.com> Date: Wed, 29 Jul 2026 15:46:41 +1000 Subject: [PATCH 63/63] chore: tweak workflow run admin to order by submission time (#114) --- app/db/admin.py | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/app/db/admin.py b/app/db/admin.py index 6326298..c5e26db 100644 --- a/app/db/admin.py +++ b/app/db/admin.py @@ -160,21 +160,22 @@ async def repr(self, obj: Any, request: Request) -> str: class WorkflowRunAdmin(ModelView): fields = [ - "id", + "submission_timestamp", + HasOne("workflow", identity="workflow"), "workflow_id", "tool", - "owner_user_id", - HasOne("workflow", identity="workflow"), HasOne("owner", identity="app-user"), + "owner_user_id", "seqera_run_id", "run_name", "service_usage", JSONField("submitted_form_data"), "binder_name", "work_dir", - "submission_timestamp", + "id", ] - exclude_fields_from_list = "submitted_form_data" + exclude_fields_from_list = ["submitted_form_data"] + fields_default_sort = [("submission_timestamp", True)] async def repr(self, obj: Any, request: Request) -> str: return f"{obj.run_name}"