Fix basecall.config path and publish .nextflow.log to LOG_BUCKET

.github/workflows/docker-build.yml

@@ -52,6 +52,6 @@ jobs:
             echo "Expected argparse exit code 2 for missing required args; got $exit_code"
             exit 1
           fi
-          for flag in --delivery --kit --aws-queue --base-bucket --work-bucket; do
+          for flag in --delivery --kit --aws-queue --base-bucket --work-bucket --log-bucket; do
             echo "$output" | grep -q -- "$flag" || { echo "Missing required flag in error output: $flag"; exit 1; }
           done

README.md

@@ -82,6 +82,7 @@ Required arguments (passed by the `startOntBasecall` Lambda via Batch `container
 - `--aws-queue` — AWS Batch GPU queue for child basecalling jobs
 - `--base-bucket` — S3 bucket holding the delivery (`raw/`, `supplemental/`, `metadata/`)
 - `--work-bucket` — S3 bucket for Nextflow's working directory
+- `--log-bucket` — S3 bucket for publishing `.nextflow.log` after the run
 
 ### Build-time authentication
 

automation/run_automation.py

@@ -18,8 +18,10 @@
 """
 
 import argparse
+import datetime as dt
 import logging
 import subprocess
+from pathlib import Path
 
 import boto3
 from botocore.config import Config
@@ -54,6 +56,10 @@ def parse_args(argv: list[str] | None = None) -> argparse.Namespace:
         "--work-bucket", required=True,
         help="S3 bucket for Nextflow's working directory",
     )
+    parser.add_argument(
+        "--log-bucket", required=True,
+        help="S3 bucket for publishing .nextflow.log after the run",
+    )
     return parser.parse_args(argv)
 
 
@@ -74,6 +80,7 @@ def build_nextflow_cmd(
     """
     return [
         "nextflow", "run", "/workflow/main.nf",
+        "-c", "/workflow/configs/basecall.config",
         "-profile", "batch",
         "--nanopore_run", delivery,
         "--kit", kit,
@@ -84,6 +91,26 @@ def build_nextflow_cmd(
     ]
 
 
+def upload_nextflow_log(
+    s3_client, delivery: str, log_bucket: str, log_path: Path = Path("/workflow/.nextflow.log"),
+) -> None:
+    """Upload .nextflow.log to s3://{log_bucket}/basecall-workflow/automated/{delivery}/{ts}/.
+
+    Logs and swallows all upload errors so it is safe to call from a finally block.
+    The error log surfaces the failure in CloudWatch so a misconfigured IAM/bucket gets noticed.
+    """
+    if not log_path.exists():
+        log.warning("No %s to upload", log_path)
+        return
+    timestamp = dt.datetime.now(dt.UTC).strftime("%Y%m%d_%H%M%S")
+    s3_key = f"basecall-workflow/automated/{delivery}/{timestamp}/.nextflow.log"
+    try:
+        s3_client.upload_file(str(log_path), log_bucket, s3_key)
+        log.info("Uploaded nextflow log to s3://%s/%s", log_bucket, s3_key)
+    except Exception as e:
+        log.exception("Failed to upload %s: %s", log_path, e)
+
+
 def main() -> None:
     logging.basicConfig(
         level=logging.INFO,
@@ -99,10 +126,17 @@ def main() -> None:
         args.work_bucket,
     )
     log.info("Running nextflow: %s", " ".join(nextflow_cmd))
-    subprocess.run(nextflow_cmd, check=True, cwd="/workflow")
+    s3_client = boto3.client("s3", config=Config(max_pool_connections=50))
+    try:
+        subprocess.run(nextflow_cmd, check=True, cwd="/workflow")
+    finally:
+        # upload_nextflow_log is designed to log-and-swallow all exceptions.
+        # (Because an exception in this finally block would prevent 
+        # samplesheet generation on a successful Nextflow run, or mask CalledProcessError 
+        # from a failed one.)
+        upload_nextflow_log(s3_client, args.delivery, args.log_bucket)
 
     log.info("Generating samplesheet for delivery %s in bucket %s", args.delivery, args.base_bucket)
-    s3_client = boto3.client("s3", config=Config(max_pool_connections=50))
     output_path = generate_samplesheet(s3_client, args.delivery, bucket=args.base_bucket)
     log.info("Samplesheet written to: %s", output_path)