Cache Playwright seed repo clones for FalkorDB test data

.github/workflows/playwright.yml

@@ -33,6 +33,20 @@ jobs:
       with:
         python-version: '3.12'
 
+    - name: Resolve seed repository SHAs
+      id: seed-repo-shas
+      run: |
+        echo "graphrag_sdk=$(git ls-remote https://github.com/FalkorDB/GraphRAG-SDK HEAD | cut -f1)" >> "$GITHUB_OUTPUT"
+        echo "flask=$(git ls-remote https://github.com/pallets/flask HEAD | cut -f1)" >> "$GITHUB_OUTPUT"
+
+    - name: Cache seeded repository clones
+      uses: actions/cache@v5
+      with:
+        path: ${{ github.workspace }}/repositories
+        key: ${{ runner.os }}-seed-repos-${{ hashFiles('e2e/seed_test_data.py') }}-${{ steps.seed-repo-shas.outputs.graphrag_sdk }}-${{ steps.seed-repo-shas.outputs.flask }}
+        restore-keys: |
+          ${{ runner.os }}-seed-repos-${{ hashFiles('e2e/seed_test_data.py') }}-
+
     - name: Install backend dependencies
       run: pip install -e .
 

e2e/seed_test_data.py

@@ -1,19 +1,22 @@
 #!/usr/bin/env python3
 """Seed FalkorDB with test data for Playwright e2e tests."""
 
-import os
-import sys
 import logging
+import os
+import shutil
+import subprocess
+from pathlib import Path
+from urllib.parse import urlparse
+
+from api.project import Project
+from falkordb import FalkorDB
 
 logging.basicConfig(
     level=logging.INFO,
     format="%(asctime)s - %(levelname)s - %(message)s",
 )
 logger = logging.getLogger(__name__)
 
-from falkordb import FalkorDB
-from api.project import Project
-
 REPOS = [
     "https://github.com/FalkorDB/GraphRAG-SDK",
     "https://github.com/pallets/flask",
@@ -25,6 +28,51 @@
     ("import_data", "add_node"),
 ]
 
+REPOSITORIES_DIR = Path(__file__).resolve().parent.parent / "repositories"
+
+
+def repo_name_from_url(url: str) -> str:
+    parsed_path = urlparse(url).path.rstrip("/")
+    repo_name = parsed_path.split("/")[-1]
+    return repo_name.removesuffix(".git")
+
+
+def fresh_clone_repository(url: str, path: Path) -> Path:
+    if path.exists():
+        # Replace any existing directory before creating a fresh shallow clone.
+        shutil.rmtree(path)
+
+    path.parent.mkdir(parents=True, exist_ok=True)
+    try:
+        subprocess.run(
+            ["git", "clone", "--depth", "1", url, str(path)],
+            check=True,
+            capture_output=True,
+            text=True,
+        )
+    except subprocess.CalledProcessError as e:
+        logger.error(
+            "git clone failed for %s (return code %s). Stderr:\n%s",
+            url,
+            e.returncode,
+            e.stderr or "<no stderr captured>",
+        )
+        raise
+
+    return path
+
+
+def load_project(url: str) -> Project:
+    repo_path = REPOSITORIES_DIR / repo_name_from_url(url)
+
+    if (repo_path / ".git").exists():
+        logger.info("Using cached repository clone at %s", repo_path)
+    else:
+        logger.info("Cloning repository into cache at %s", repo_path)
+        fresh_clone_repository(url, repo_path)
+
+    return Project.from_local_repository(repo_path)
+
 
 def ensure_calls_edges(graph_name: str) -> None:
     """Ensure required CALLS edges exist for E2E tests.
@@ -63,7 +111,7 @@ def ensure_calls_edges(graph_name: str) -> None:
 def main():
     for url in REPOS:
         logger.info("Seeding %s ...", url)
-        proj = Project.from_git_repository(url)
+        proj = load_project(url)
         proj.analyze_sources()
         logger.info("Done seeding %s", url)
 

tests/test_seed_test_data.py

@@ -0,0 +1,51 @@
+from pathlib import Path
+
+import e2e.seed_test_data as seed_test_data
+
+
+def test_load_project_uses_existing_cached_clone(monkeypatch, tmp_path):
+    repo_path = tmp_path / "GraphRAG-SDK"
+    (repo_path / ".git").mkdir(parents=True)
+
+    calls = []
+
+    class FakeProject:
+        @staticmethod
+        def from_local_repository(path):
+            calls.append(path)
+            return path
+
+    monkeypatch.setattr(seed_test_data, "REPOSITORIES_DIR", tmp_path)
+    monkeypatch.setattr(seed_test_data, "Project", FakeProject)
+
+    project = seed_test_data.load_project("https://github.com/FalkorDB/GraphRAG-SDK")
+
+    assert project == repo_path
+    assert calls == [repo_path]
+
+
+def test_load_project_clones_into_cache(monkeypatch, tmp_path):
+    repo_path = tmp_path / "flask"
+    clone_calls = []
+    project_calls = []
+
+    class FakeProject:
+        @staticmethod
+        def from_local_repository(path):
+            project_calls.append(path)
+            return path
+
+    def fake_clone(url: str, path: Path) -> Path:
+        clone_calls.append((url, path))
+        (path / ".git").mkdir(parents=True)
+        return path
+
+    monkeypatch.setattr(seed_test_data, "REPOSITORIES_DIR", tmp_path)
+    monkeypatch.setattr(seed_test_data, "Project", FakeProject)
+    monkeypatch.setattr(seed_test_data, "fresh_clone_repository", fake_clone)
+
+    project = seed_test_data.load_project("https://github.com/pallets/flask")
+
+    assert project == repo_path
+    assert clone_calls == [("https://github.com/pallets/flask", repo_path)]
+    assert project_calls == [repo_path]