ci: prepare regression datasets before test execution and verify immutability

Author: harryswift01

.github/workflows/pr.yaml

@@ -110,37 +110,39 @@ jobs:
         shell: bash
         run: |
           rm -rf .pytest_cache
-          rm -rf .testdata
           find . -type d -name "__pycache__" -prune -exec rm -rf {} +
 
-      - name: Snapshot regression test data before tests
+      - name: Prepare regression test data
         shell: bash
         run: |
-          if [ -d .testdata ]; then
-            find .testdata -type f -exec sha256sum {} \; | sort > testdata.before.sha256
-          else
-            touch testdata.before.sha256
+          python -m tests.regression.prepare_testdata --system "${{ matrix.system }}"
+
+          if [ ! -d .testdata ]; then
+            echo ".testdata was not created"
+            exit 1
           fi
 
+      - name: Snapshot regression test data before tests
+        shell: bash
+        run: |
+          find .testdata -type f -exec sha256sum {} \; | sort > testdata.before.sha256
+
       - name: Run fast regression tests (per system)
         run: |
           python -m pytest tests/regression \
             -m "not slow" \
             --cache-clear \
             -n 1 \
-            -k "${{ matrix.system }}" \
+            --system "${{ matrix.system }}" \
             -vv \
-            --durations=20
+            --durations=20 \
+            --codeentropy-debug
 
       - name: Snapshot regression test data after tests
         if: always()
         shell: bash
         run: |
-          if [ -d .testdata ]; then
-            find .testdata -type f -exec sha256sum {} \; | sort > testdata.after.sha256
-          else
-            touch testdata.after.sha256
-          fi
+          find .testdata -type f -exec sha256sum {} \; | sort > testdata.after.sha256
 
       - name: Check regression test data did not change
         if: always()

tests/regression/conftest.py

@@ -1,6 +1,5 @@
 from __future__ import annotations
 
-import os
 import random
 
 import numpy as np
@@ -48,8 +47,6 @@ def pytest_configure(config: pytest.Config) -> None:
     random.seed(seed)
     np.random.seed(seed)
 
-    os.environ["PYTHONHASHSEED"] = "0"
-
 
 def pytest_collection_modifyitems(
     config: pytest.Config, items: list[pytest.Item]

tests/regression/prepare_testdata.py

@@ -0,0 +1,54 @@
+from __future__ import annotations
+
+import argparse
+from pathlib import Path
+
+import yaml
+
+from tests.regression.cases import discover_cases
+from tests.regression.helpers import (
+    _abspathify_config_paths,
+    ensure_testdata_for_system,
+)
+
+
+def required_paths_for_case(case) -> list[Path]:
+    raw = yaml.safe_load(case.config_path.read_text())
+    cooked = _abspathify_config_paths(raw, base_dir=case.config_path.parent)
+
+    required: list[Path] = []
+    run1 = cooked.get("run1")
+
+    if isinstance(run1, dict):
+        ff = run1.get("force_file")
+        if isinstance(ff, str) and ff:
+            required.append(Path(ff))
+
+        for p in run1.get("top_traj_file") or []:
+            if isinstance(p, str) and p:
+                required.append(Path(p))
+
+    return required
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--system", action="append", default=None)
+    args = parser.parse_args()
+
+    selected = set(args.system or [])
+    cases = discover_cases()
+
+    for case in cases:
+        if selected and case.system not in selected:
+            continue
+
+        required = required_paths_for_case(case)
+        if required:
+            ensure_testdata_for_system(case.system, required_paths=required)
+
+    print("Regression test data prepared.")
+
+
+if __name__ == "__main__":
+    main()