"""Tests for feature extraction and scoring."""
import pytest
from fed_speech_mcp.features import FeatureExtractor, ImportanceScorer
from fed_speech_mcp.models import (
DocType,
Features,
ImportanceTier,
Speaker,
SpeakerRole,
TopicFlags,
)
class TestFeatureExtractor:
"""Tests for FeatureExtractor."""
@pytest.fixture
def extractor(self):
"""Create a FeatureExtractor."""
return FeatureExtractor()
def test_word_count(self, extractor):
"""Test word count extraction."""
text = "This is a simple test sentence with ten words total."
features = extractor.extract(text)
assert features.word_count == 10
def test_topic_detection_inflation(self, extractor):
"""Test inflation topic detection."""
text = "The inflation rate continues to be elevated. PCE remains above target."
features = extractor.extract(text)
assert features.topics.inflation is True
def test_topic_detection_rates(self, extractor):
"""Test rates topic detection."""
text = "The federal funds rate will remain elevated. Monetary policy is restrictive."
features = extractor.extract(text)
assert features.topics.rates is True
def test_topic_detection_labor(self, extractor):
"""Test labor market topic detection."""
text = "The labor market remains tight. Unemployment is low and wages are rising."
features = extractor.extract(text)
assert features.topics.labor_market is True
def test_topic_detection_multiple(self, extractor):
"""Test detecting multiple topics."""
text = """
Inflation remains above our 2% target while the labor market
continues to show strength. We will maintain restrictive monetary policy
and keep interest rates elevated.
"""
features = extractor.extract(text)
assert features.topics.inflation is True
assert features.topics.labor_market is True
assert features.topics.rates is True
def test_qa_detection(self, extractor):
"""Test Q&A section detection."""
text = "Thank you for the speech. Now let's open it up for Q&A. Q: What about rates?"
features = extractor.extract(text)
assert features.has_qa is True
def test_no_qa_detection(self, extractor):
"""Test no Q&A detection when not present."""
text = "This is a prepared speech without any questions."
features = extractor.extract(text)
assert features.has_qa is False
def test_language_default(self, extractor):
"""Test default language is English."""
features = extractor.extract("Any text")
assert features.language == "en"
class TestImportanceScorer:
"""Tests for ImportanceScorer."""
@pytest.fixture
def scorer(self):
"""Create an ImportanceScorer."""
return ImportanceScorer()
def test_chair_base_tier(self, scorer):
"""Test Chair gets high base tier."""
speaker = Speaker(name="Jerome Powell", role=SpeakerRole.CHAIR)
features = Features(word_count=1000, topics=TopicFlags())
importance = scorer.score(speaker, DocType.SPEECH, features)
assert importance.tier == ImportanceTier.HIGH
def test_governor_base_tier(self, scorer):
"""Test Governor gets medium base tier."""
speaker = Speaker(name="Michelle Bowman", role=SpeakerRole.GOVERNOR)
features = Features(word_count=1000, topics=TopicFlags())
importance = scorer.score(speaker, DocType.SPEECH, features)
assert importance.tier == ImportanceTier.MEDIUM
def test_testimony_boost(self, scorer):
"""Test testimony increases importance."""
speaker = Speaker(name="Michelle Bowman", role=SpeakerRole.GOVERNOR)
features = Features(word_count=1000, topics=TopicFlags())
speech_importance = scorer.score(speaker, DocType.SPEECH, features)
testimony_importance = scorer.score(speaker, DocType.TESTIMONY, features)
assert testimony_importance.score > speech_importance.score
def test_qa_boost(self, scorer):
"""Test Q&A presence increases importance."""
speaker = Speaker(name="Michelle Bowman", role=SpeakerRole.GOVERNOR)
features_no_qa = Features(word_count=1000, has_qa=False, topics=TopicFlags())
features_qa = Features(word_count=1000, has_qa=True, topics=TopicFlags())
importance_no_qa = scorer.score(speaker, DocType.SPEECH, features_no_qa)
importance_qa = scorer.score(speaker, DocType.SPEECH, features_qa)
assert importance_qa.score > importance_no_qa.score
def test_topic_boost(self, scorer):
"""Test rates + inflation/labor topics increase importance."""
speaker = Speaker(name="Michelle Bowman", role=SpeakerRole.GOVERNOR)
features_no_topics = Features(
word_count=1000,
topics=TopicFlags(),
)
features_with_topics = Features(
word_count=1000,
topics=TopicFlags(rates=True, inflation=True),
)
importance_no = scorer.score(speaker, DocType.SPEECH, features_no_topics)
importance_with = scorer.score(speaker, DocType.SPEECH, features_with_topics)
assert importance_with.score > importance_no.score
def test_short_document_penalty(self, scorer):
"""Test short documents get reduced importance."""
speaker = Speaker(name="Jerome Powell", role=SpeakerRole.CHAIR)
features_short = Features(word_count=100, topics=TopicFlags())
features_normal = Features(word_count=1000, topics=TopicFlags())
importance_short = scorer.score(speaker, DocType.SPEECH, features_short)
importance_normal = scorer.score(speaker, DocType.SPEECH, features_normal)
assert importance_short.score < importance_normal.score
def test_reasons_populated(self, scorer):
"""Test that reasons are populated."""
speaker = Speaker(name="Jerome Powell", role=SpeakerRole.CHAIR)
features = Features(
word_count=1000,
has_qa=True,
topics=TopicFlags(rates=True, inflation=True),
)
importance = scorer.score(speaker, DocType.TESTIMONY, features)
assert len(importance.reasons) >= 3
assert any("Chair" in r for r in importance.reasons)
def test_score_bounds(self, scorer):
"""Test score is always between 0 and 1."""
speaker = Speaker(name="Test", role=SpeakerRole.GOVERNOR)
features = Features(word_count=50, topics=TopicFlags())
importance = scorer.score(speaker, DocType.SPEECH, features)
assert 0 <= importance.score <= 1
