feat(fc3c): GalleryDLService — subprocess wrap + error classification + validate/quarantine

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-20 20:41:42 -04:00
parent 0c80c12984
commit 7a96b1d06c
2 changed files with 811 additions and 0 deletions
@@ -0,0 +1,151 @@
+"""GalleryDLService tests — subprocess is mocked so no real gallery-dl runs."""
+
+from types import SimpleNamespace
+
+import pytest
+
+from backend.app.services.gallery_dl import (
+    ErrorType,
+    GalleryDLService,
+    SourceConfig,
+)
+
+
+@pytest.fixture
+def gdl(tmp_path):
+    return GalleryDLService(
+        images_root=tmp_path / "images",
+        rate_limit=3.0,
+        validate_files=False,
+    )
+
+
+def _proc(stdout="", stderr="", returncode=0):
+    return SimpleNamespace(stdout=stdout, stderr=stderr, returncode=returncode)
+
+
+@pytest.mark.asyncio
+async def test_download_happy_path(gdl, monkeypatch):
+    out_path = gdl.images_root / "alice" / "patreon" / "post1" / "img.jpg"
+    out_path.parent.mkdir(parents=True, exist_ok=True)
+    out_path.write_bytes(b"\xff\xd8\xff\xe0" + b"\x00" * 32 + b"\xff\xd9")
+
+    monkeypatch.setattr(
+        "backend.app.services.gallery_dl.subprocess.run",
+        lambda *a, **k: _proc(stdout=str(out_path) + "\n", returncode=0),
+    )
+    result = await gdl.download(
+        url="https://patreon.com/alice",
+        artist_slug="alice",
+        platform="patreon",
+        source_config=SourceConfig(),
+    )
+    assert result.success is True
+    assert result.files_downloaded == 1
+    assert result.error_type is None
+    assert str(out_path) in result.written_paths
+
+
+@pytest.mark.asyncio
+async def test_download_no_new_content(gdl, monkeypatch):
+    monkeypatch.setattr(
+        "backend.app.services.gallery_dl.subprocess.run",
+        lambda *a, **k: _proc(stdout="# /images/alice/patreon/old.jpg\n", returncode=0),
+    )
+    result = await gdl.download(
+        url="https://patreon.com/alice", artist_slug="alice", platform="patreon",
+        source_config=SourceConfig(),
+    )
+    assert result.success is True
+    assert result.files_downloaded == 0
+
+
+@pytest.mark.asyncio
+async def test_download_auth_error_categorization(gdl, monkeypatch):
+    monkeypatch.setattr(
+        "backend.app.services.gallery_dl.subprocess.run",
+        lambda *a, **k: _proc(
+            stdout="", stderr='[patreon][error] "GET /api/foo HTTP/1.1" 401 None\n',
+            returncode=1,
+        ),
+    )
+    result = await gdl.download(
+        url="https://patreon.com/alice", artist_slug="alice", platform="patreon",
+        source_config=SourceConfig(),
+    )
+    assert result.success is False
+    assert result.error_type == ErrorType.AUTH_ERROR
+
+
+@pytest.mark.asyncio
+async def test_download_timeout(gdl, monkeypatch):
+    import subprocess as sp
+
+    def _raise(*a, **k):
+        raise sp.TimeoutExpired(cmd="gallery-dl", timeout=10)
+
+    monkeypatch.setattr("backend.app.services.gallery_dl.subprocess.run", _raise)
+    result = await gdl.download(
+        url="https://patreon.com/alice", artist_slug="alice", platform="patreon",
+        source_config=SourceConfig(timeout=10),
+    )
+    assert result.success is False
+    assert result.error_type == ErrorType.TIMEOUT
+
+
+@pytest.mark.asyncio
+async def test_validation_quarantines_truncated_files(tmp_path, monkeypatch):
+    gdl = GalleryDLService(
+        images_root=tmp_path / "images", rate_limit=3.0, validate_files=True,
+    )
+    bad = gdl.images_root / "alice" / "patreon" / "post" / "bad.jpg"
+    bad.parent.mkdir(parents=True, exist_ok=True)
+    bad.write_bytes(b"\xff\xd8\xff" + b"\x00" * 100)  # SOI but no EOI
+
+    monkeypatch.setattr(
+        "backend.app.services.gallery_dl.subprocess.run",
+        lambda *a, **k: _proc(stdout=str(bad) + "\n", returncode=0),
+    )
+    result = await gdl.download(
+        url="https://patreon.com/alice", artist_slug="alice", platform="patreon",
+        source_config=SourceConfig(),
+    )
+    assert result.files_quarantined == 1
+    assert len(result.quarantined_paths) == 1
+    assert not bad.exists()
+    assert result.error_type == ErrorType.VALIDATION_FAILED
+
+
+def test_compute_run_stats(gdl):
+    stats = gdl._compute_run_stats(
+        return_code=0,
+        stdout="/path/a.jpg\n/path/b.jpg\n# /path/c.jpg\n",
+        stderr="[patreon][warning] foo\n[patreon][download][error] failed to download /x\n",
+    )
+    assert stats["downloaded_count"] == 2
+    assert stats["skipped_count"] == 1
+    assert stats["warning_count"] >= 1
+    assert stats["per_item_failures"] == 1
+
+
+def test_extract_errors_warnings(gdl):
+    stderr = (
+        "[patreon][debug] foo\n[patreon][error] bar\n"
+        "[urllib3][debug] baz\n[patreon][warning] qux\n"
+    )
+    result = gdl._extract_errors_warnings(stderr)
+    assert "[error]" in result
+    assert "[warning]" in result
+    assert "[debug]" not in result
+
+
+def test_truncate_log_caps_large_text(gdl):
+    huge = "line\n" * 200_000
+    result = gdl._truncate_log(huge, max_bytes=10_000)
+    assert len(result.encode()) < 11_000
+    assert "lines elided" in result
+
+
+def test_truncate_log_passes_short_text(gdl):
+    text = "small\n"
+    assert gdl._truncate_log(text) == text