Skip to content

Codebase intelligence for AI and humans: code health scores, auto-generated docs, git analytics, dead code detection, and architectural decisions via MCP.

test: run pipeline tests on a tmp copy of sample_repo (#4091) (#4099)

Dhvani Dave3586e5c0179d
diff --git a/tests/conftest.py b/tests/conftest.py
index ba7bba0c9..b7eebab26 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -3,6 +3,8 @@
from __future__ import annotations
import os
+import shutil
+import subprocess
import sys
from pathlib import Path
@@ -249,6 +251,26 @@ def sample_repo_path(repo_root: Path) -> Path:
return path
+def sample_repo_copy(sample_repo_path: Path, tmp_path: Path) -> Path:
+ """A git-backed copy of sample_repo for tests that write into it.
+
+ The pipeline reads git history, so the copy gets one commit of its own.
+ """
+ dest = tmp_path / "sample_repo"
+ shutil.copytree(sample_repo_path, dest)
+ env = {
+ **os.environ,
+ "GIT_AUTHOR_NAME": "t",
+ "GIT_AUTHOR_EMAIL": "[email protected]",
+ "GIT_COMMITTER_NAME": "t",
+ "GIT_COMMITTER_EMAIL": "[email protected]",
+ }
+ for args in (["init", "-q"], ["add", "-A"], ["commit", "-q", "-m", "init"]):
+ subprocess.run(["git", *args], cwd=dest, check=True, capture_output=True, env=env)
+ return dest
+
+
@pytest.fixture(scope="session")
def fixtures_dir(repo_root: Path) -> Path:
"""Path to the tests/fixtures/ directory."""
diff --git a/tests/integration/conftest.py b/tests/integration/conftest.py
new file mode 100644
index 000000000..b371795c7
--- /dev/null
+++ b/tests/integration/conftest.py
@@ -0,0 +1,18 @@
+"""Shared setup for the integration tests."""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+import pytest
+
+
[email protected](scope="session", autouse=True)
+def sample_repo_stays_clean(sample_repo_path: Path):
+ """Fail the run if a test wrote .repowise/ into the shared sample repo."""
+ leftover = sample_repo_path / ".repowise"
+ yield
+ assert not leftover.exists(), (
+ f"{leftover} was created during the run. Use the sample_repo_copy "
+ "fixture for tests that write into the sample repo."
+ )
diff --git a/tests/integration/test_pipeline_fast_checkpoint.py b/tests/integration/test_pipeline_fast_checkpoint.py
index 585304fbd..0cb56a686 100644
--- a/tests/integration/test_pipeline_fast_checkpoint.py
+++ b/tests/integration/test_pipeline_fast_checkpoint.py
@@ -55,10 +55,10 @@ class _FakeJobStore:
return matches[:limit]
-async def test_fast_mode_skips_generation_and_co_change(sample_repo_path: Path) -> None:
+async def test_fast_mode_skips_generation_and_co_change(sample_repo_copy: Path) -> None:
store = _FakeJobStore()
result = await run_pipeline(
- sample_repo_path,
+ sample_repo_copy,
generate_docs=True, # FAST must override this to False (no LLM client passed)
mode=OrchestratorMode.FAST,
job_store=store,
@@ -72,15 +72,15 @@ async def test_fast_mode_skips_generation_and_co_change(sample_repo_path: Path)
assert meta.get("co_change_partners_json", "[]") == "[]"
-async def test_checkpoints_recorded_for_completed_phases(sample_repo_path: Path) -> None:
+async def test_checkpoints_recorded_for_completed_phases(sample_repo_copy: Path) -> None:
store = _FakeJobStore()
await run_pipeline(
- sample_repo_path,
+ sample_repo_copy,
mode=OrchestratorMode.FAST,
job_store=store,
)
- completed = await find_completed_phases(store, str(Path(sample_repo_path).resolve()))
+ completed = await find_completed_phases(store, str(Path(sample_repo_copy).resolve()))
# parse + git always run; both should be recorded COMPLETED.
assert "parse" in completed
assert "git" in completed
diff --git a/tests/integration/test_pipeline_resume.py b/tests/integration/test_pipeline_resume.py
index b4694e47a..c187dd584 100644
--- a/tests/integration/test_pipeline_resume.py
+++ b/tests/integration/test_pipeline_resume.py
@@ -42,12 +42,12 @@ async def _make_repo(sf, path: Path) -> str:
return repo.id
-async def test_first_run_checkpoints_index(sf, sample_repo_path: Path) -> None:
- repo_id = await _make_repo(sf, sample_repo_path)
+async def test_first_run_checkpoints_index(sf, sample_repo_copy: Path) -> None:
+ repo_id = await _make_repo(sf, sample_repo_copy)
ctrl = ResumeController(sf, repo_id, resume=False)
result = await run_pipeline(
- sample_repo_path,
+ sample_repo_copy,
mode=OrchestratorMode.FAST,
resume_controller=ctrl,
)
@@ -58,12 +58,12 @@ async def test_first_run_checkpoints_index(sf, sample_repo_path: Path) -> None:
assert ResumePhase.INDEX in completed
-async def test_resume_skips_index_compute(sf, sample_repo_path: Path, monkeypatch) -> None:
- repo_id = await _make_repo(sf, sample_repo_path)
+async def test_resume_skips_index_compute(sf, sample_repo_copy: Path, monkeypatch) -> None:
+ repo_id = await _make_repo(sf, sample_repo_copy)
# First run persists the INDEX phase.
await run_pipeline(
- sample_repo_path,
+ sample_repo_copy,
mode=OrchestratorMode.FAST,
resume_controller=ResumeController(sf, repo_id, resume=False),
)
@@ -82,7 +82,7 @@ async def test_resume_skips_index_compute(sf, sample_repo_path: Path, monkeypatc
monkeypatch.setattr(orch, "_run_git_indexing", _boom_git)
result = await run_pipeline(
- sample_repo_path,
+ sample_repo_copy,
mode=OrchestratorMode.FAST,
resume_controller=ResumeController(sf, repo_id, resume=True),
)
@@ -93,10 +93,10 @@ async def test_resume_skips_index_compute(sf, sample_repo_path: Path, monkeypatc
assert result.graph_builder.graph().number_of_nodes() > 0, "graph should be rehydrated"
-async def test_first_run_checkpoints_analysis(sf, sample_repo_path: Path) -> None:
- repo_id = await _make_repo(sf, sample_repo_path)
+async def test_first_run_checkpoints_analysis(sf, sample_repo_copy: Path) -> None:
+ repo_id = await _make_repo(sf, sample_repo_copy)
await run_pipeline(
- sample_repo_path,
+ sample_repo_copy,
mode=OrchestratorMode.FAST,
resume_controller=ResumeController(sf, repo_id, resume=False),
)
@@ -107,12 +107,12 @@ async def test_first_run_checkpoints_analysis(sf, sample_repo_path: Path) -> Non
assert ResumePhase.ANALYSIS in completed
-async def test_resume_skips_analysis_recompute(sf, sample_repo_path: Path, monkeypatch) -> None:
- repo_id = await _make_repo(sf, sample_repo_path)
+async def test_resume_skips_analysis_recompute(sf, sample_repo_copy: Path, monkeypatch) -> None:
+ repo_id = await _make_repo(sf, sample_repo_copy)
# First run persists INDEX + ANALYSIS.
await run_pipeline(
- sample_repo_path,
+ sample_repo_copy,
mode=OrchestratorMode.FAST,
resume_controller=ResumeController(sf, repo_id, resume=False),
)
@@ -135,7 +135,7 @@ async def test_resume_skips_analysis_recompute(sf, sample_repo_path: Path, monke
monkeypatch.setattr(orch, "_run_decision_extraction", _boom_decisions)
result = await run_pipeline(
- sample_repo_path,
+ sample_repo_copy,
mode=OrchestratorMode.FAST,
resume_controller=ResumeController(sf, repo_id, resume=True),
)
@@ -143,14 +143,14 @@ async def test_resume_skips_analysis_recompute(sf, sample_repo_path: Path, monke
async def test_analysis_checkpoint_receives_the_runs_vector_store(
- sf, sample_repo_path: Path, monkeypatch
+ sf, sample_repo_copy: Path, monkeypatch
) -> None:
"""The store a caller hands ``run_pipeline`` has to reach the checkpoint.
Dropping it there is invisible: the run still succeeds and still embeds, it
just stores a duplicate instead of folding it.
"""
- repo_id = await _make_repo(sf, sample_repo_path)
+ repo_id = await _make_repo(sf, sample_repo_copy)
seen: list[object] = []
async def _capture(_self: object, **kwargs: object) -> None:
@@ -160,7 +160,7 @@ async def test_analysis_checkpoint_receives_the_runs_vector_store(
store = object()
await run_pipeline(
- sample_repo_path,
+ sample_repo_copy,
vector_store=store,
resume_controller=ResumeController(sf, repo_id, resume=False),
)
diff --git a/tests/integration/test_upgrade_fast_to_full.py b/tests/integration/test_upgrade_fast_to_full.py
index d5b0a225a..1db6cb457 100644
--- a/tests/integration/test_upgrade_fast_to_full.py
+++ b/tests/integration/test_upgrade_fast_to_full.py
@@ -53,16 +53,16 @@ async def session():
async def test_fast_index_then_full_upgrade(
- sample_repo_path: Path,
+ sample_repo_copy: Path,
session: AsyncSession,
tmp_path: Path,
monkeypatch: pytest.MonkeyPatch,
) -> None:
# --- 1. Fast index + persist (ESSENTIAL git, no docs). ---
- repo = await upsert_repository(session, name="sample", local_path=str(sample_repo_path))
+ repo = await upsert_repository(session, name="sample", local_path=str(sample_repo_copy))
repo_id = repo.id
- result = await run_pipeline(sample_repo_path, mode=OrchestratorMode.FAST)
+ result = await run_pipeline(sample_repo_copy, mode=OrchestratorMode.FAST)
assert result.generated_pages is None # FAST never generates docs
assert result.git_metadata_list, "expected git metadata for the sample repo"
assert all(
@@ -89,14 +89,14 @@ async def test_fast_index_then_full_upgrade(
monkeypatch.setattr(GraphBuilder, "build", _spy_build)
# --- 3. Rehydrate the graph from SQL — equivalence + zero resolution. ---
- rehydrated = await rehydrate_graph_builder(session, repo_id, sample_repo_path)
+ rehydrated = await rehydrate_graph_builder(session, repo_id, sample_repo_copy)
assert rehydrated.graph().number_of_nodes() == original_nodes
assert rehydrated.graph().number_of_edges() == original_edges
assert rehydrated.pagerank() == original_pagerank
assert build_calls["n"] == 0, "rehydration must not re-resolve the graph"
# --- 4. Backfill the git tier ESSENTIAL -> FULL (resumable via JobStore). ---
- indexer = GitIndexer(sample_repo_path, tier=GitIndexTier.FULL)
+ indexer = GitIndexer(sample_repo_copy, tier=GitIndexTier.FULL)
job_store = SqlJobStore(session)
summary, git_results = await backfill_full_tier(indexer, repo_id, job_store=job_store)
await session.commit()