feat(benchmark): add database query benchmark infrastructure

.github/workflows/benchmark.yml

@@ -0,0 +1,58 @@
+name: "Benchmark"
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+jobs:
+  benchmark:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v6
+        with:
+          ref: ${{ github.event.workflow_run.head_branch }}
+          token: ${{ github.actor == 'dependabot[bot]' && secrets.REPO_AND_READ_PACKAGES_PAT || secrets.XML_SUBMODULE_PAT }}
+          submodules: "recursive"
+
+      - name: Set up Python 3.14
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.14"
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install -e ".[sqlmodels,ahbicht,benchmark]"
+
+      - name: Download baseline (if PR)
+        id: download-baseline
+        if: github.event_name == 'pull_request'
+        uses: dawidd6/action-download-artifact@v6
+        with:
+          name: benchmark-baseline
+          branch: main
+          path: .benchmarks/baseline
+          if_no_artifact_found: warn
+
+      - name: Run benchmarks
+        run: |
+          pytest benchmarks/ --benchmark-only --benchmark-json=benchmark-results.json
+
+      - name: Compare against baseline (if PR and baseline exists)
+        if: github.event_name == 'pull_request'
+        run: |
+          if [ -f ".benchmarks/baseline/benchmark-results.json" ]; then
+            python scripts/compare_benchmarks.py .benchmarks/baseline/benchmark-results.json benchmark-results.json
+          else
+            echo "No baseline found. Skipping comparison (first run or expired artifact)."
+          fi
+
+      - name: Save baseline (if main)
+        if: github.ref == 'refs/heads/main'
+        uses: actions/upload-artifact@v4
+        with:
+          name: benchmark-baseline
+          path: benchmark-results.json
+          retention-days: 90

benchmarks/__init__.py

@@ -0,0 +1 @@
+# Benchmark tests for database query performance

benchmarks/conftest.py

@@ -0,0 +1,15 @@
+# Re-export fixtures from unittests for use in benchmarks
+# The database creation happens in the fixture (module-scoped), not during benchmark timing.
+# Only the SELECT queries inside benchmark() are timed.
+
+from unittests.conftest import (
+    is_private_submodule_checked_out,
+    session_fv2410_fv2504_with_diff_view,
+    session_fv2510_fv2604_mscons_with_diff_view,
+)
+
+__all__ = [
+    "is_private_submodule_checked_out",
+    "session_fv2410_fv2504_with_diff_view",
+    "session_fv2510_fv2604_mscons_with_diff_view",
+]

benchmarks/test_diff_query.py

@@ -0,0 +1,58 @@
+"""
+Benchmarks for v_ahb_diff query performance.
+
+These benchmarks measure SELECT query time only - database creation happens
+in the module-scoped fixtures and is not included in the timing.
+
+Run with: pytest benchmarks/ --benchmark-only
+Save results: pytest benchmarks/ --benchmark-only --benchmark-json=results.json
+"""
+
+import pytest
+from sqlmodel import Session, text
+
+# FV2410 -> FV2504 benchmarks with representative pruefidentifikators
+PRUEFIS_FV2410_FV2504 = ["55109", "13002", "25001"]
+
+
+@pytest.mark.benchmark(group="v_ahb_diff_fv2410_fv2504")
+@pytest.mark.parametrize("pruefi", PRUEFIS_FV2410_FV2504)
+def test_diff_query_fv2410_fv2504(benchmark, session_fv2410_fv2504_with_diff_view: Session, pruefi: str) -> None:
+    """Benchmark diff query: FV2410 -> FV2504 with same pruefidentifikator."""
+    query = text(
+        """
+        SELECT * FROM v_ahb_diff
+        WHERE new_pruefidentifikator = :pruefi
+          AND old_pruefidentifikator = :pruefi
+          AND new_format_version = 'FV2504'
+          AND old_format_version = 'FV2410'
+        ORDER BY sort_path ASC
+    """
+    )
+
+    def run_query():
+        return list(session_fv2410_fv2504_with_diff_view.execute(query, {"pruefi": pruefi}))
+
+    result = benchmark.pedantic(run_query, iterations=10, rounds=1)
+    assert len(result) > 0, f"Query returned no results for pruefi {pruefi}"
+
+
+@pytest.mark.benchmark(group="v_ahb_diff_fv2510_fv2604")
+def test_diff_query_fv2510_fv2604_13009(benchmark, session_fv2510_fv2604_mscons_with_diff_view: Session) -> None:
+    """Benchmark diff query: FV2510 -> FV2604, pruefidentifikator 13009."""
+    query = text(
+        """
+        SELECT * FROM v_ahb_diff
+        WHERE new_pruefidentifikator = '13009'
+          AND old_pruefidentifikator = '13009'
+          AND new_format_version = 'FV2510'
+          AND old_format_version = 'FV2604'
+        ORDER BY sort_path ASC
+    """
+    )
+
+    def run_query():
+        return list(session_fv2510_fv2604_mscons_with_diff_view.execute(query, {"pruefi": "13009"}))
+
+    result = benchmark.pedantic(run_query, iterations=10, rounds=1)
+    assert len(result) > 0, "Query returned no results for pruefi 13009"

pyproject.toml

@@ -56,6 +56,10 @@ tests = [
     "pytest==9.0.2",
     "syrupy==5.0.0"
 ]
+benchmark = [
+    "pytest==9.0.2",
+    "pytest-benchmark==5.1.0"
+]
 type_check = [
     "mypy==1.19.1"
 ]

scripts/compare_benchmarks.py

@@ -0,0 +1,86 @@
+#!/usr/bin/env python3
+"""Compare benchmark results and report regressions."""
+import argparse
+import json
+import sys
+from pathlib import Path
+
+REGRESSION_THRESHOLD = 0.20  # 20% slower = regression
+
+
+def load_benchmarks(path: Path) -> dict[str, float]:
+    """Load benchmark results and return dict of test name -> median time."""
+    with open(path) as f:
+        data = json.load(f)
+    return {b["name"]: b["stats"]["median"] for b in data["benchmarks"]}
+
+
+def main() -> int:
+    parser = argparse.ArgumentParser(description="Compare benchmark results and report regressions.")
+    parser.add_argument("baseline", type=Path, help="Path to baseline benchmark JSON file")
+    parser.add_argument("current", type=Path, help="Path to current benchmark JSON file")
+    args = parser.parse_args()
+
+    baseline_path: Path = args.baseline
+    current_path: Path = args.current
+
+    if not baseline_path.exists():
+        print("No baseline found, skipping comparison.")
+        return 0
+
+    try:
+        baseline = load_benchmarks(baseline_path)
+    except (json.JSONDecodeError, KeyError) as e:
+        print(f"Error reading baseline: {e}")
+        return 0
+
+    try:
+        current = load_benchmarks(current_path)
+    except (json.JSONDecodeError, KeyError) as e:
+        print(f"Error reading current results: {e}")
+        return 1
+
+    regressions: list[tuple[str, float]] = []
+    print("\n## Benchmark Comparison\n")
+    print("| Test | Baseline | Current | Change |")
+    print("|------|----------|---------|--------|")
+
+    for name in sorted(set(baseline.keys()) | set(current.keys())):
+        if name not in current:
+            print(f"| {name} | {baseline[name]*1000:.1f}ms | - | REMOVED |")
+            continue
+
+        current_time = current[name]
+
+        if name not in baseline:
+            print(f"| {name} | - | {current_time*1000:.1f}ms | NEW |")
+            continue
+
+        baseline_time = baseline[name]
+        change = (current_time - baseline_time) / baseline_time
+
+        if change > REGRESSION_THRESHOLD:
+            status = "🔴"
+        elif change < -0.05:
+            status = "🟢"
+        else:
+            status = ""
+
+        print(f"| {name} | {baseline_time*1000:.1f}ms | {current_time*1000:.1f}ms | {change:+.1%} {status} |")
+
+        if change > REGRESSION_THRESHOLD:
+            regressions.append((name, change))
+
+    print()
+    if regressions:
+        print(f"❌ {len(regressions)} regression(s) detected (>{REGRESSION_THRESHOLD:.0%} slower)")
+        for name, change in regressions:
+            print(f"  - {name}: {change:+.1%}")
+        return 1
+    else:
+        print("✅ No regressions detected")
+        return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())