Improve EverCore Python quality gates

.github/workflows/evercore-smoke.yml

@@ -53,6 +53,30 @@ jobs:
           PYTHONPATH: src:.
         run: uv run pytest tests/test_content_item_compat.py -q
 
+      # Phase 2 T2.1 — runs pytest in CI. Marker-filtered to "unit" via the
+      # path heuristic in tests/conftest.py. Many existing tests still
+      # require databases despite that classification; keep
+      # continue-on-error until tests are explicitly tagged per T2.4.
+      - name: Run unit tests with coverage (baseline; non-blocking)
+        env:
+          PYTHONPATH: src:.
+        run: uv run pytest tests/ -m unit --cov=src --cov-report=xml --cov-report=term --maxfail=20
+        continue-on-error: true
+
+      - name: Upload coverage artifact
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: evercore-coverage
+          path: methods/EverCore/coverage.xml
+          if-no-files-found: warn
+
+      - name: Type check (baseline; warnings only)
+        env:
+          PYTHONPATH: src:.
+        run: uv run ty check --ignore all --error possibly-unresolved-reference
+        continue-on-error: true
+
       - name: Validate demo payload compatibility
         env:
           PYTHONPATH: src:.

.gitignore

@@ -209,6 +209,9 @@ AGENTS.mk
 docs/api_docs/profile_extraction_fields.md
 .claude/
 .cursor/*
+# Per-developer CLAUDE.md overrides (loaded by Claude Code alongside CLAUDE.md)
+CLAUDE.local.md
+**/CLAUDE.local.md
 
 #tmp_data
 demo/memcell_outputs/

methods/EverCore/Makefile

@@ -1,14 +1,23 @@
-.PHONY: dev-setup setup-hooks lint test clean help
+.PHONY: dev-setup setup-hooks lint ruff ruff-fix typecheck typecheck-pyright test test-unit test-integration test-e2e test-cov benchmark clean help
 
 # Default target
 help:
 	@echo "Available targets:"
-	@echo "  dev-setup        - Full dev environment setup (sync deps + install hooks)"
-	@echo "  setup-hooks      - Install pre-commit hooks only"
-	@echo "  lint             - Run linters"
-	@echo "  test             - Run tests"
-	@echo "  clean            - Clean up generated files"
-	@echo "  help             - Show this help message"
+	@echo "  dev-setup          - Full dev environment setup (sync deps + install hooks)"
+	@echo "  setup-hooks        - Install pre-commit hooks only"
+	@echo "  lint               - Run linters (black + i18n)"
+	@echo "  ruff               - Run ruff check + format check (read-only)"
+	@echo "  ruff-fix           - Run ruff with --fix and apply formatter"
+	@echo "  typecheck          - Run ty type checker (default, fast)"
+	@echo "  typecheck-pyright  - Run pyright type checker (fallback)"
+	@echo "  test               - Run all tests"
+	@echo "  test-unit          - Run unit tests only (no docker required)"
+	@echo "  test-integration   - Run integration tests (needs docker-compose up)"
+	@echo "  test-e2e           - Run end-to-end tests"
+	@echo "  test-cov           - Run unit tests with coverage report (xml + term)"
+	@echo "  benchmark          - Run perf benchmarks (needs bench group + docker)"
+	@echo "  clean              - Clean up generated files"
+	@echo "  help               - Show this help message"
 
 # Full development environment setup
 dev-setup:
@@ -34,9 +43,63 @@ lint:
 	@echo "Running i18n check..."
 	PYTHONPATH=src python -m devops_scripts.i18n.i18n_tool check
 
+# Run ruff (lint + format check). Read-only; CI-friendly.
+ruff:
+	@echo "Running ruff check..."
+	uv run ruff check src/
+	@echo "Running ruff format --check..."
+	uv run ruff format --check src/
+
+# Apply ruff auto-fixes and formatting.
+ruff-fix:
+	@echo "Running ruff check --fix..."
+	uv run ruff check --fix src/
+	@echo "Running ruff format..."
+	uv run ruff format src/
+
+# Run ty type checker (default).
+# Configured in pyproject.toml under [tool.ty]. ~10x faster than pyright.
+# Note: ty is pre-1.0 — pin a floor in pyproject.toml and accept that diagnostic
+# wording may shift between minor releases.
+typecheck:
+	@echo "Running ty..."
+	uv run ty check --ignore all --error possibly-unresolved-reference
+
+# Run pyright type checker (fallback).
+# pyrightconfig.json controls scope and strictness; current mode is "off"
+# (syntax/import resolution only). Tighten there to surface more checks.
+typecheck-pyright:
+	@echo "Running pyright..."
+	uv run pyright
+
 # Run tests
 test:
-	PYTHONPATH=src pytest tests/
+	PYTHONPATH=src uv run pytest tests/
+
+# Unit tests only. No docker-compose dependencies; safe for CI on every PR.
+# Markers are applied automatically in tests/conftest.py based on file path.
+test-unit:
+	PYTHONPATH=src uv run pytest tests/ -m unit
+
+# Integration tests. Require docker-compose services (Redis, MongoDB, ES, Milvus).
+test-integration:
+	PYTHONPATH=src uv run pytest tests/ -m integration
+
+# End-to-end tests. Heaviest tier; run on demand or in a nightly job.
+test-e2e:
+	PYTHONPATH=src uv run pytest tests/ -m e2e
+
+# Coverage report (XML for CI tooling + terminal summary).
+test-cov:
+	PYTHONPATH=src uv run pytest tests/ -m unit \
+		--cov=src --cov-report=xml --cov-report=term
+
+# Performance benchmarks. Requires the optional `bench` group
+# (`uv sync --group bench`) and RUN_BENCHMARKS=1 plus a running
+# docker-compose stack. See tests/benchmarks/README.md.
+benchmark:
+	PYTHONPATH=src RUN_BENCHMARKS=1 uv run pytest tests/benchmarks/ \
+		--benchmark-only --benchmark-columns=min,mean,median,p95,max
 
 # Clean up
 clean: