Files
veritext/tests/test_metrics/test_readability.py

275 lines
10 KiB
Python

"""Tests for the readability metric."""
import pytest
from veritext.metrics import Readability, ReadabilityResult
class TestReadability:
"""Tests for the Readability metric class."""
@pytest.fixture
def readability(self) -> Readability:
"""Provide a readability metric instance."""
return Readability()
def test_name(self, readability: Readability) -> None:
"""Test that name returns 'readability'."""
assert readability.name == "readability"
def test_requires_reference(self, readability: Readability) -> None:
"""Test that readability does NOT require reference text."""
assert readability.requires_reference is False
def test_simple_text(self, readability: Readability) -> None:
"""Test readability of simple, easy text."""
# Simple children's text - short sentences, simple words
text = "The cat sat. The dog ran. I see a bird."
result = readability.score(text)
# Should have low grade level and high reading ease
assert result.flesch_kincaid_grade < 5.0
assert result.flesch_reading_ease > 80.0
def test_complex_text(self, readability: Readability) -> None:
"""Test readability of complex, academic text."""
# Complex academic text - long sentences, polysyllabic words
text = (
"The implementation of sophisticated computational methodologies "
"necessitates comprehensive understanding of algorithmic complexity "
"and architectural considerations."
)
result = readability.score(text)
# Should have high grade level and low reading ease
assert result.flesch_kincaid_grade > 12.0
assert result.flesch_reading_ease < 30.0
def test_medium_text(self, readability: Readability) -> None:
"""Test readability of medium-difficulty text."""
text = (
"The weather today is quite pleasant. "
"Many people are enjoying the sunshine in the park. "
"Children play while parents watch nearby."
)
result = readability.score(text)
# Should be middle of the road
assert 3.0 < result.flesch_kincaid_grade < 10.0
assert 50.0 < result.flesch_reading_ease < 90.0
def test_single_sentence(self, readability: Readability) -> None:
"""Test readability with a single sentence."""
text = "The cat sat on the mat."
result = readability.score(text)
# Should compute without error
assert result.flesch_kincaid_grade is not None
assert result.flesch_reading_ease is not None
def test_single_word(self, readability: Readability) -> None:
"""Test readability with a single word."""
text = "Cat"
result = readability.score(text)
# Should handle single word (1 word, 1 sentence, 1 syllable)
assert result.flesch_kincaid_grade is not None
assert result.flesch_reading_ease is not None
def test_empty_text(self, readability: Readability) -> None:
"""Test that empty text returns zero scores."""
result = readability.score("")
assert result.flesch_kincaid_grade == 0.0
assert result.flesch_reading_ease == 0.0
def test_whitespace_only(self, readability: Readability) -> None:
"""Test that whitespace-only text returns zero scores."""
result = readability.score(" \t\n ")
assert result.flesch_kincaid_grade == 0.0
assert result.flesch_reading_ease == 0.0
def test_reference_ignored(self, readability: Readability) -> None:
"""Test that reference parameter is ignored."""
text = "The cat sat on the mat."
# Score with no reference
result1 = readability.score(text)
# Score with reference (should be ignored)
result2 = readability.score(text, "Completely different text")
# Score with list of references
result3 = readability.score(text, ["ref1", "ref2"])
# All should produce identical results
assert result1.flesch_kincaid_grade == result2.flesch_kincaid_grade
assert result1.flesch_reading_ease == result2.flesch_reading_ease
assert result1.flesch_kincaid_grade == result3.flesch_kincaid_grade
def test_punctuation_handling(self, readability: Readability) -> None:
"""Test that punctuation affects sentence counting."""
# Same words, different sentence structure
text1 = "The cat sat on the mat" # 1 sentence
text2 = "The cat sat. On the mat." # 2 sentences
result1 = readability.score(text1)
result2 = readability.score(text2)
# Different sentence counts should affect scores
assert result1.flesch_kincaid_grade != result2.flesch_kincaid_grade
def test_question_marks_count_sentences(self, readability: Readability) -> None:
"""Test that question marks end sentences."""
text = "What is this? It is a test."
result = readability.score(text)
# Should count as 2 sentences
# With 7 words total, words_per_sentence = 3.5
assert result.flesch_kincaid_grade is not None
def test_exclamation_marks_count_sentences(self, readability: Readability) -> None:
"""Test that exclamation marks end sentences."""
text = "Wow! That is amazing!"
result = readability.score(text)
# Should count as 2 sentences
assert result.flesch_kincaid_grade is not None
def test_multiple_punctuation(self, readability: Readability) -> None:
"""Test handling of multiple punctuation marks."""
text = "What?! That's crazy... Well then."
result = readability.score(text)
# Should handle gracefully
assert result.flesch_kincaid_grade is not None
def test_result_score_property(self, readability: Readability) -> None:
"""Test that result.score returns flesch_reading_ease."""
result = readability.score("The cat sat on the mat.")
assert result.score == result.flesch_reading_ease
def test_contractions(self, readability: Readability) -> None:
"""Test handling of contractions."""
text = "I'm going to the store. It's not far away."
result = readability.score(text)
# Should handle contractions as words
assert result.flesch_kincaid_grade is not None
assert result.flesch_reading_ease is not None
class TestReadabilityBatch:
"""Tests for readability batch scoring."""
@pytest.fixture
def readability(self) -> Readability:
"""Provide a readability metric instance."""
return Readability()
def test_batch_score_basic(self, readability: Readability) -> None:
"""Test basic batch scoring."""
candidates = [
"The cat sat on the mat.",
"A dog ran through the park.",
]
result = readability.batch_score(candidates)
assert result.count == 2
assert len(result.results) == 2
def test_batch_score_statistics(self, readability: Readability) -> None:
"""Test that batch scoring computes statistics."""
candidates = [
"Cat sat.", # Very simple
"The implementation of sophisticated methodologies requires expertise.",
]
result = readability.batch_score(candidates)
# Check statistics are computed
assert "flesch_kincaid_grade" in result.stats
assert "flesch_reading_ease" in result.stats
# First should be easier than second
assert (
result.results[0].flesch_reading_ease
> result.results[1].flesch_reading_ease
)
def test_batch_score_percentiles(self, readability: Readability) -> None:
"""Test that batch scoring computes percentiles."""
candidates = ["a", "b", "c", "d", "e"]
result = readability.batch_score(candidates)
stats = result.stats["flesch_reading_ease"]
assert 25 in stats.percentiles
assert 50 in stats.percentiles
assert 75 in stats.percentiles
assert 95 in stats.percentiles
def test_batch_score_references_ignored(self, readability: Readability) -> None:
"""Test that batch scoring ignores references."""
candidates = ["The cat sat.", "A dog ran."]
result1 = readability.batch_score(candidates)
result2 = readability.batch_score(candidates, ["ref1", "ref2"])
# Results should be identical
assert result1.results[0].flesch_kincaid_grade == (
result2.results[0].flesch_kincaid_grade
)
def test_batch_score_empty_list_raises(self, readability: Readability) -> None:
"""Test that empty candidate list raises ValueError."""
with pytest.raises(ValueError, match="empty"):
readability.batch_score([])
class TestReadabilityResult:
"""Tests for ReadabilityResult type."""
def test_frozen(self) -> None:
"""Test that ReadabilityResult is frozen."""
from pydantic import ValidationError
result = ReadabilityResult(flesch_kincaid_grade=5.0, flesch_reading_ease=70.0)
with pytest.raises(ValidationError):
result.flesch_kincaid_grade = 6.0 # type: ignore[misc]
def test_values(self) -> None:
"""Test that values are stored correctly."""
result = ReadabilityResult(flesch_kincaid_grade=8.5, flesch_reading_ease=65.0)
assert result.flesch_kincaid_grade == 8.5
assert result.flesch_reading_ease == 65.0
def test_score_property(self) -> None:
"""Test that score property returns flesch_reading_ease."""
result = ReadabilityResult(flesch_kincaid_grade=8.5, flesch_reading_ease=65.0)
assert result.score == 65.0
class TestSyllableCounting:
"""Tests for syllable counting heuristics."""
@pytest.fixture
def readability(self) -> Readability:
"""Provide a readability metric instance."""
return Readability()
def test_monosyllabic_words(self, readability: Readability) -> None:
"""Test that monosyllabic words don't inflate scores."""
# All one-syllable words
text = "The cat sat on the mat."
result = readability.score(text)
# Should be very easy to read
assert result.flesch_reading_ease > 90.0
def test_polysyllabic_words(self, readability: Readability) -> None:
"""Test that polysyllabic words affect scores."""
# Words with multiple syllables
text = "International communication facilitates understanding."
result = readability.score(text)
# Should be harder to read
assert result.flesch_reading_ease < 50.0