ComfyUI/tests-unit/assets_test/services/test_enrich.py
Luke Mino-Altherr 196959472a Clean up imports, remove section banners, and fix seed endpoint with autoscan disabled
Move inline/lazy imports to top-level now that blake3 is a main dependency and
circular-import concerns are resolved. Remove decorative section-header comments
in asset_reference queries. Allow the seed API endpoint to temporarily re-enable
the seeder when --disable-assets-autoscan is active so manual triggers still work.
Add retry-on-409 logic in test helpers and remove try/finally around prompt
execution in favour of sequential pause/resume.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-24 11:34:44 -08:00

253 lines
7.7 KiB
Python

"""Tests for asset enrichment (mime_type and hash population)."""
from pathlib import Path
from unittest.mock import patch
from sqlalchemy.orm import Session
from app.assets.database.models import Asset, AssetReference
from app.assets.scanner import (
ENRICHMENT_HASHED,
ENRICHMENT_METADATA,
ENRICHMENT_STUB,
enrich_asset,
)
def _create_stub_asset(
session: Session,
file_path: str,
asset_id: str = "test-asset-id",
reference_id: str = "test-ref-id",
name: str | None = None,
) -> tuple[Asset, AssetReference]:
"""Create a stub asset with reference for testing enrichment."""
asset = Asset(
id=asset_id,
hash=None,
size_bytes=100,
mime_type=None,
)
session.add(asset)
session.flush()
ref = AssetReference(
id=reference_id,
asset_id=asset_id,
name=name or f"test-asset-{asset_id}",
owner_id="system",
file_path=file_path,
mtime_ns=1234567890000000000,
enrichment_level=ENRICHMENT_STUB,
)
session.add(ref)
session.flush()
return asset, ref
class TestEnrichAsset:
def test_extracts_mime_type_and_updates_asset(
self, db_engine, temp_dir: Path, session: Session
):
"""Verify mime_type is written to the Asset table during enrichment."""
file_path = temp_dir / "model.safetensors"
file_path.write_bytes(b"\x00" * 100)
asset, ref = _create_stub_asset(
session, str(file_path), "asset-1", "ref-1"
)
session.commit()
with patch("app.assets.scanner.create_session") as mock_cs:
from contextlib import contextmanager
@contextmanager
def _create_session():
with Session(db_engine) as sess:
yield sess
mock_cs.side_effect = _create_session
new_level = enrich_asset(
file_path=str(file_path),
reference_id=ref.id,
asset_id=asset.id,
extract_metadata=True,
compute_hash=False,
)
assert new_level == ENRICHMENT_METADATA
session.expire_all()
updated_asset = session.get(Asset, "asset-1")
assert updated_asset is not None
assert updated_asset.mime_type == "application/safetensors"
def test_computes_hash_and_updates_asset(
self, db_engine, temp_dir: Path, session: Session
):
"""Verify hash is written to the Asset table during enrichment."""
file_path = temp_dir / "data.bin"
file_path.write_bytes(b"test content for hashing")
asset, ref = _create_stub_asset(
session, str(file_path), "asset-2", "ref-2"
)
session.commit()
with patch("app.assets.scanner.create_session") as mock_cs:
from contextlib import contextmanager
@contextmanager
def _create_session():
with Session(db_engine) as sess:
yield sess
mock_cs.side_effect = _create_session
new_level = enrich_asset(
file_path=str(file_path),
reference_id=ref.id,
asset_id=asset.id,
extract_metadata=True,
compute_hash=True,
)
assert new_level == ENRICHMENT_HASHED
session.expire_all()
updated_asset = session.get(Asset, "asset-2")
assert updated_asset is not None
assert updated_asset.hash is not None
assert updated_asset.hash.startswith("blake3:")
def test_enrichment_updates_both_mime_and_hash(
self, db_engine, temp_dir: Path, session: Session
):
"""Verify both mime_type and hash are set when full enrichment runs."""
file_path = temp_dir / "model.safetensors"
file_path.write_bytes(b"\x00" * 50)
asset, ref = _create_stub_asset(
session, str(file_path), "asset-3", "ref-3"
)
session.commit()
with patch("app.assets.scanner.create_session") as mock_cs:
from contextlib import contextmanager
@contextmanager
def _create_session():
with Session(db_engine) as sess:
yield sess
mock_cs.side_effect = _create_session
enrich_asset(
file_path=str(file_path),
reference_id=ref.id,
asset_id=asset.id,
extract_metadata=True,
compute_hash=True,
)
session.expire_all()
updated_asset = session.get(Asset, "asset-3")
assert updated_asset is not None
assert updated_asset.mime_type == "application/safetensors"
assert updated_asset.hash is not None
assert updated_asset.hash.startswith("blake3:")
def test_missing_file_returns_stub_level(
self, db_engine, temp_dir: Path, session: Session
):
"""Verify missing files don't cause errors and return STUB level."""
file_path = temp_dir / "nonexistent.bin"
asset, ref = _create_stub_asset(
session, str(file_path), "asset-4", "ref-4"
)
session.commit()
with patch("app.assets.scanner.create_session") as mock_cs:
from contextlib import contextmanager
@contextmanager
def _create_session():
with Session(db_engine) as sess:
yield sess
mock_cs.side_effect = _create_session
new_level = enrich_asset(
file_path=str(file_path),
reference_id=ref.id,
asset_id=asset.id,
extract_metadata=True,
compute_hash=True,
)
assert new_level == ENRICHMENT_STUB
session.expire_all()
updated_asset = session.get(Asset, "asset-4")
assert updated_asset.mime_type is None
assert updated_asset.hash is None
def test_duplicate_hash_merges_into_existing_asset(
self, db_engine, temp_dir: Path, session: Session
):
"""Verify duplicate files merge into existing asset instead of failing."""
file_path_1 = temp_dir / "file1.bin"
file_path_2 = temp_dir / "file2.bin"
content = b"identical content"
file_path_1.write_bytes(content)
file_path_2.write_bytes(content)
asset1, ref1 = _create_stub_asset(
session, str(file_path_1), "asset-dup-1", "ref-dup-1"
)
asset2, ref2 = _create_stub_asset(
session, str(file_path_2), "asset-dup-2", "ref-dup-2"
)
session.commit()
with patch("app.assets.scanner.create_session") as mock_cs:
from contextlib import contextmanager
@contextmanager
def _create_session():
with Session(db_engine) as sess:
yield sess
mock_cs.side_effect = _create_session
enrich_asset(
file_path=str(file_path_1),
reference_id=ref1.id,
asset_id=asset1.id,
extract_metadata=True,
compute_hash=True,
)
enrich_asset(
file_path=str(file_path_2),
reference_id=ref2.id,
asset_id=asset2.id,
extract_metadata=True,
compute_hash=True,
)
session.expire_all()
updated_asset1 = session.get(Asset, "asset-dup-1")
assert updated_asset1 is not None
assert updated_asset1.hash is not None
updated_asset2 = session.get(Asset, "asset-dup-2")
assert updated_asset2 is None
updated_ref2 = session.get(AssetReference, "ref-dup-2")
assert updated_ref2 is not None
assert updated_ref2.asset_id == "asset-dup-1"