Apply Linux analysis process exit fix #2
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: Apply deterministic postprocessing CLI exit fix | |
| on: | |
| push: | |
| branches: | |
| - agent/release-readiness-audit | |
| permissions: | |
| contents: write | |
| jobs: | |
| apply-fix: | |
| if: github.actor != 'github-actions[bot]' | |
| runs-on: ubuntu-latest | |
| steps: | |
| - uses: actions/checkout@v4 | |
| with: | |
| ref: agent/release-readiness-audit | |
| fetch-depth: 0 | |
| - uses: actions/setup-python@v5 | |
| with: | |
| python-version: '3.11' | |
| - name: Install package and test tooling | |
| run: | | |
| python -m pip install --upgrade pip | |
| python -m pip install --no-cache-dir -e . pytest | |
| - name: Apply deterministic CLI and wrapper fix | |
| shell: bash | |
| run: | | |
| set -euo pipefail | |
| python - <<'PY' | |
| from pathlib import Path | |
| cli = Path('src/wwgpt/cli.py') | |
| text = cli.read_text() | |
| seed_marker = '''def _seeds(s: str | None) -> list[int] | None: | |
| return None if not s else [int(x) for x in s.split(',') if x] | |
| ''' | |
| seed_replacement = '''def _seeds(s: str | None) -> list[int] | None: | |
| return None if not s else [int(x) for x in s.split(',') if x] | |
| def _exit_after_flush(status: int = 0) -> None: | |
| """Exit a completed CLI command without waiting on library worker threads.""" | |
| sys.stderr.flush() | |
| sys.stdout.flush() | |
| os._exit(status) | |
| ''' | |
| if 'def _exit_after_flush(status: int = 0)' not in text: | |
| if seed_marker not in text: | |
| raise SystemExit('CLI seed helper marker not found') | |
| text = text.replace(seed_marker, seed_replacement, 1) | |
| old_analyze = ' elif args.cmd=="analyze-results": print(analyze_results(args.results_root, args.analysis_plan))\n' | |
| new_analyze = ''' elif args.cmd=="analyze-results": | |
| print(analyze_results(args.results_root, args.analysis_plan), flush=True) | |
| _exit_after_flush(0) | |
| ''' | |
| if old_analyze in text: | |
| text = text.replace(old_analyze, new_analyze, 1) | |
| elif new_analyze.strip() not in text: | |
| raise SystemExit('analyze-results CLI branch not found') | |
| old_prepare = ''' sys.stderr.flush() | |
| sys.stdout.flush() | |
| # Some streaming dataset backends can leave non-daemon workers alive after all artifacts | |
| # have been written. Exit the CLI process deterministically so shell wrappers can finish. | |
| os._exit(0) | |
| ''' | |
| new_prepare = ''' # Some data and analysis backends can leave non-daemon workers alive after all | |
| # artifacts have been written. Exit deterministically so shell wrappers finish. | |
| _exit_after_flush(0) | |
| ''' | |
| if old_prepare in text: | |
| text = text.replace(old_prepare, new_prepare, 1) | |
| old_report = ''' elif args.cmd=="generate-reproducibility-report": | |
| print( | |
| write_reproducibility_report( | |
| args.experiment_root, | |
| strict=args.strict, | |
| analysis_plan=args.analysis_plan, | |
| ) | |
| ) | |
| ''' | |
| new_report = ''' elif args.cmd=="generate-reproducibility-report": | |
| print( | |
| write_reproducibility_report( | |
| args.experiment_root, | |
| strict=args.strict, | |
| analysis_plan=args.analysis_plan, | |
| ), | |
| flush=True, | |
| ) | |
| _exit_after_flush(0) | |
| ''' | |
| if old_report in text: | |
| text = text.replace(old_report, new_report, 1) | |
| elif new_report.strip() not in text: | |
| raise SystemExit('reproducibility CLI branch not found') | |
| cli.write_text(text) | |
| release_test = Path('tests/test_release_readiness_audit.py') | |
| tests = release_test.read_text() | |
| test_block = ''' | |
| def test_postprocessing_cli_exits_after_flushing_completed_artifacts() -> None: | |
| source = Path("src/wwgpt/cli.py").read_text() | |
| assert "def _exit_after_flush(status: int = 0)" in source | |
| assert 'elif args.cmd=="analyze-results":\\n print(' in source | |
| assert 'elif args.cmd=="generate-reproducibility-report":' in source | |
| assert source.count("_exit_after_flush(0)") >= 3 | |
| ''' | |
| if 'test_postprocessing_cli_exits_after_flushing_completed_artifacts' not in tests: | |
| release_test.write_text(tests + test_block) | |
| local_test = Path('tests/test_local_level_scripts.py') | |
| tests = local_test.read_text() | |
| local_block = ''' | |
| def test_bounded_level012_runner_normalizes_report_output_and_emits_diagnostics() -> None: | |
| source = Path("scripts/run_bounded_level012_acceptance.sh").read_text() | |
| assert "set -Eeuo pipefail" in source | |
| assert "trap on_error ERR" in source | |
| assert "extract_report_path" in source | |
| assert "release_reproducibility_first.log" in source | |
| assert "release_reproducibility_second.log" in source | |
| assert 'FIRST_REPORT="$(extract_report_path "$FIRST_LOG")"' in source | |
| assert 'SECOND_REPORT="$(extract_report_path "$SECOND_LOG")"' in source | |
| assert 'FIRST_REPORT="$(wwgpt generate-reproducibility-report' not in source | |
| ''' | |
| if 'test_bounded_level012_runner_normalizes_report_output_and_emits_diagnostics' not in tests: | |
| local_test.write_text(tests + local_block) | |
| PY | |
| cat > scripts/run_bounded_level012_acceptance.sh <<'EOF' | |
| #!/usr/bin/env bash | |
| set -Eeuo pipefail | |
| if [ "$#" -ne 1 ]; then | |
| echo "usage: $0 RESULTS_ROOT" >&2 | |
| exit 2 | |
| fi | |
| SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" | |
| REPO_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)" | |
| cd "$REPO_ROOT" | |
| RESULTS_ROOT="$(python - "$1" <<'PY' | |
| from pathlib import Path | |
| import sys | |
| print(Path(sys.argv[1]).expanduser().resolve()) | |
| PY | |
| )" | |
| PLAN="$RESULTS_ROOT/release_acceptance_plan.yaml" | |
| on_error() { | |
| local status=$? | |
| local line=${BASH_LINENO[0]:-unknown} | |
| local command=${BASH_COMMAND:-unknown} | |
| echo "[release-acceptance] failure status=$status line=$line command=$command" >&2 | |
| for log in \ | |
| "$RESULTS_ROOT/release_health.log" \ | |
| "$RESULTS_ROOT/release_analysis.log" \ | |
| "$RESULTS_ROOT/release_audit.log" \ | |
| "$RESULTS_ROOT/release_reproducibility_first.log" \ | |
| "$RESULTS_ROOT/release_reproducibility_second.log"; do | |
| if [ -f "$log" ]; then | |
| echo "[release-acceptance] tail $log" >&2 | |
| tail -n 40 "$log" >&2 || true | |
| fi | |
| done | |
| exit "$status" | |
| } | |
| trap on_error ERR | |
| extract_report_path() { | |
| python - "$1" <<'PY' | |
| from pathlib import Path | |
| import sys | |
| log = Path(sys.argv[1]) | |
| lines = [line.strip() for line in log.read_text().splitlines() if line.strip()] | |
| candidates = [line for line in lines if line.endswith(".pdf")] | |
| if not candidates: | |
| raise SystemExit(f"no PDF report path found in {log}") | |
| report = Path(candidates[-1]).expanduser() | |
| if not report.is_absolute(): | |
| report = (Path.cwd() / report).resolve() | |
| else: | |
| report = report.resolve() | |
| if not report.is_file(): | |
| raise SystemExit(f"reported PDF does not exist: {report}") | |
| print(report) | |
| PY | |
| } | |
| python scripts/run_bounded_level012_acceptance.py --results-root "$RESULTS_ROOT" | |
| wwgpt check-health --experiment-root "$RESULTS_ROOT" >"$RESULTS_ROOT/release_health.log" | |
| wwgpt analyze-results "$RESULTS_ROOT" --analysis-plan "$PLAN" >"$RESULTS_ROOT/release_analysis.log" | |
| wwgpt audit-experiment --experiment-root "$RESULTS_ROOT" >"$RESULTS_ROOT/release_audit.log" | |
| FIRST_LOG="$RESULTS_ROOT/release_reproducibility_first.log" | |
| SECOND_LOG="$RESULTS_ROOT/release_reproducibility_second.log" | |
| wwgpt generate-reproducibility-report \ | |
| --experiment-root "$RESULTS_ROOT" \ | |
| --analysis-plan "$PLAN" \ | |
| --strict >"$FIRST_LOG" | |
| wwgpt generate-reproducibility-report \ | |
| --experiment-root "$RESULTS_ROOT" \ | |
| --analysis-plan "$PLAN" \ | |
| --strict >"$SECOND_LOG" | |
| FIRST_REPORT="$(extract_report_path "$FIRST_LOG")" | |
| SECOND_REPORT="$(extract_report_path "$SECOND_LOG")" | |
| if [ "$FIRST_REPORT" != "$SECOND_REPORT" ]; then | |
| echo "reproducibility report path changed across identical reruns" >&2 | |
| echo "first: $FIRST_REPORT" >&2 | |
| echo "second: $SECOND_REPORT" >&2 | |
| exit 1 | |
| fi | |
| python - "$RESULTS_ROOT" <<'PY' | |
| from pathlib import Path | |
| import json | |
| import sys | |
| import pandas as pd | |
| root = Path(sys.argv[1]).resolve() | |
| analysis = root / "analysis" | |
| complete = sorted(root.rglob("run_complete.json")) | |
| if len(complete) != 6: | |
| raise SystemExit(f"expected six complete runs, found {len(complete)}") | |
| inventory = pd.read_csv(analysis / "runs_manifest.csv") | |
| if len(inventory) != 6: | |
| raise SystemExit(f"expected six analyzed arms, found {len(inventory)}") | |
| if set(pd.to_numeric(inventory["level"], errors="raise").astype(int)) != {0, 1, 2}: | |
| raise SystemExit("analysis did not preserve Levels 0, 1, and 2") | |
| required = [ | |
| analysis / "analysis_eligibility.json", | |
| analysis / "acceleration_by_seed.csv", | |
| analysis / "integrity_summary.json", | |
| analysis / "reproducibility_report.json", | |
| analysis / "reproducibility_report.pdf", | |
| analysis / "cross_level_run_inventory.csv", | |
| ] | |
| missing = [str(path) for path in required if not path.is_file()] | |
| if missing: | |
| raise SystemExit(f"missing release artifacts: {missing}") | |
| eligibility = json.loads((analysis / "analysis_eligibility.json").read_text()) | |
| if not eligibility.get("eligible"): | |
| raise SystemExit(f"analysis ineligible: {eligibility}") | |
| print("LEVEL_0_1_2_RELEASE_ACCEPTANCE_PASS") | |
| PY | |
| EOF | |
| chmod +x scripts/run_bounded_level012_acceptance.sh | |
| rm -f .github/workflows/apply-linux-cli-exit-fix.yml | |
| - name: Validate focused and clean-process paths | |
| env: | |
| MPLBACKEND: Agg | |
| run: | | |
| set -euo pipefail | |
| python -m compileall -q src tests scripts | |
| bash -n scripts/*.sh | |
| pytest -q \ | |
| tests/test_release_readiness_audit.py \ | |
| tests/test_local_level_scripts.py \ | |
| tests/test_acceleration_analysis.py \ | |
| tests/test_level012_full_pipeline.py | |
| timeout 300 ./scripts/run_bounded_level012_acceptance.sh /tmp/linux-cli-exit-acceptance | |
| - name: Commit validated fix | |
| shell: bash | |
| run: | | |
| set -euo pipefail | |
| git config user.name github-actions[bot] | |
| git config user.email 41898282+github-actions[bot]@users.noreply.github.com | |
| git add src/wwgpt/cli.py scripts/run_bounded_level012_acceptance.sh \ | |
| tests/test_release_readiness_audit.py tests/test_local_level_scripts.py \ | |
| .github/workflows/apply-linux-cli-exit-fix.yml | |
| git diff --cached --check | |
| git commit -m 'Exit completed postprocessing commands deterministically' | |
| git push origin HEAD:agent/release-readiness-audit |