mirror of
https://github.com/nomadkaraoke/karaoke-gen.git
synced 2026-11-20 13:07:57 +09:00
## Summary - Preload SpaCy `en_core_web_sm` model at FastAPI container startup to eliminate 60+ second delay during first agentic correction job - Add `backend/services/spacy_preloader.py` singleton service for model storage - Modify `PhraseAnalyzer` and `SyllablesMatchHandler` to use preloaded model when available - Add timing logs to verify optimization is working ## Background Job `2ccbdf6b` investigation revealed a 63-second gap in logs caused by SpaCy model loading during `AnchorSequenceFinder` initialization. Cloud Run's filesystem I/O is slower than local SSD, making lazy loading expensive on first request. ## Changes | File | Change | |------|--------| | `backend/services/spacy_preloader.py` | New singleton service for preloading | | `backend/main.py` | Call preload in FastAPI lifespan handler | | `phrase_analyzer.py` | Check for preloaded model first | | `syllables_match.py` | Check for preloaded model first | | `anchor_sequence.py` | Add timing logs to initialization | | `backend/tests/test_spacy_preloader.py` | Unit tests (7 tests) | ## Expected Logging (when working) **Startup:** ``` Preloading SpaCy model 'en_core_web_sm'... SpaCy model 'en_core_web_sm' preloaded in X.XXs ``` **During job:** ``` Initializing AnchorSequenceFinder... Using preloaded SpaCy model: en_core_web_sm Initialized AnchorSequenceFinder in 0.XXs ``` ## Test plan - [x] Unit tests pass (7/7 for spacy_preloader) - [x] Full test suite passes (1018 passed, 2 pre-existing failures unrelated) - [ ] Deploy and run test job to verify logs show preloaded model usage - [ ] Verify `Initialized AnchorSequenceFinder` time is < 1 second <!-- @coderabbitai ignore --> 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
120 lines
4.3 KiB
Python
120 lines
4.3 KiB
Python
"""Tests for SpaCy preloader service."""
|
|
|
|
import pytest
|
|
from unittest.mock import patch, MagicMock
|
|
|
|
from backend.services.spacy_preloader import (
|
|
preload_spacy_model,
|
|
get_preloaded_model,
|
|
is_model_preloaded,
|
|
clear_preloaded_models,
|
|
)
|
|
|
|
|
|
class TestSpacyPreloader:
|
|
"""Tests for SpaCy preloading functionality."""
|
|
|
|
def setup_method(self):
|
|
"""Clear preloaded models before each test."""
|
|
clear_preloaded_models()
|
|
|
|
def teardown_method(self):
|
|
"""Clear preloaded models after each test."""
|
|
clear_preloaded_models()
|
|
|
|
def test_preload_spacy_model_loads_and_stores(self):
|
|
"""GIVEN no preloaded models
|
|
WHEN preload_spacy_model is called
|
|
THEN model should be loaded and stored in singleton."""
|
|
mock_nlp = MagicMock()
|
|
|
|
with patch("spacy.load", return_value=mock_nlp) as mock_load:
|
|
preload_spacy_model("en_core_web_sm")
|
|
|
|
mock_load.assert_called_once_with("en_core_web_sm")
|
|
assert is_model_preloaded("en_core_web_sm")
|
|
assert get_preloaded_model("en_core_web_sm") is mock_nlp
|
|
|
|
def test_preload_is_idempotent(self):
|
|
"""GIVEN a model already preloaded
|
|
WHEN preload_spacy_model is called again
|
|
THEN model should not be reloaded."""
|
|
mock_nlp = MagicMock()
|
|
|
|
with patch("spacy.load", return_value=mock_nlp) as mock_load:
|
|
preload_spacy_model("en_core_web_sm")
|
|
preload_spacy_model("en_core_web_sm") # Second call
|
|
|
|
# Should only load once
|
|
assert mock_load.call_count == 1
|
|
|
|
def test_get_preloaded_model_returns_none_if_not_loaded(self):
|
|
"""GIVEN no preloaded models
|
|
WHEN get_preloaded_model is called
|
|
THEN should return None."""
|
|
assert get_preloaded_model("en_core_web_sm") is None
|
|
assert not is_model_preloaded("en_core_web_sm")
|
|
|
|
def test_preload_different_models(self):
|
|
"""GIVEN no preloaded models
|
|
WHEN preload_spacy_model is called with different model names
|
|
THEN each model should be loaded and stored separately."""
|
|
mock_nlp_sm = MagicMock(name="en_core_web_sm")
|
|
mock_nlp_md = MagicMock(name="en_core_web_md")
|
|
|
|
def mock_load(model_name):
|
|
if model_name == "en_core_web_sm":
|
|
return mock_nlp_sm
|
|
elif model_name == "en_core_web_md":
|
|
return mock_nlp_md
|
|
raise ValueError(f"Unknown model: {model_name}")
|
|
|
|
with patch("spacy.load", side_effect=mock_load):
|
|
preload_spacy_model("en_core_web_sm")
|
|
preload_spacy_model("en_core_web_md")
|
|
|
|
assert is_model_preloaded("en_core_web_sm")
|
|
assert is_model_preloaded("en_core_web_md")
|
|
assert get_preloaded_model("en_core_web_sm") is mock_nlp_sm
|
|
assert get_preloaded_model("en_core_web_md") is mock_nlp_md
|
|
|
|
def test_clear_preloaded_models(self):
|
|
"""GIVEN preloaded models
|
|
WHEN clear_preloaded_models is called
|
|
THEN all models should be removed."""
|
|
mock_nlp = MagicMock()
|
|
|
|
with patch("spacy.load", return_value=mock_nlp):
|
|
preload_spacy_model("en_core_web_sm")
|
|
assert is_model_preloaded("en_core_web_sm")
|
|
|
|
clear_preloaded_models()
|
|
assert not is_model_preloaded("en_core_web_sm")
|
|
assert get_preloaded_model("en_core_web_sm") is None
|
|
|
|
def test_preload_failure_raises_exception(self):
|
|
"""GIVEN spacy.load raises an exception
|
|
WHEN preload_spacy_model is called
|
|
THEN exception should be propagated."""
|
|
with patch(
|
|
"spacy.load",
|
|
side_effect=OSError("Model not found"),
|
|
):
|
|
with pytest.raises(OSError, match="Model not found"):
|
|
preload_spacy_model("nonexistent_model")
|
|
|
|
# Model should not be marked as preloaded
|
|
assert not is_model_preloaded("nonexistent_model")
|
|
|
|
def test_preload_uses_default_model_name(self):
|
|
"""GIVEN no model name specified
|
|
WHEN preload_spacy_model is called without arguments
|
|
THEN should use en_core_web_sm as default."""
|
|
mock_nlp = MagicMock()
|
|
|
|
with patch("spacy.load", return_value=mock_nlp) as mock_load:
|
|
preload_spacy_model()
|
|
|
|
mock_load.assert_called_once_with("en_core_web_sm")
|
|
assert is_model_preloaded("en_core_web_sm")
|