mirror of
https://github.com/open-thought/reasoning-gym.git
synced 2026-04-19 12:58:07 +00:00
103 lines
3.4 KiB
Python
103 lines
3.4 KiB
Python
"""Tests for letter counting task generation"""
|
|
|
|
import pytest
|
|
|
|
from reasoning_gym.algorithmic.letter_counting import (
|
|
LetterCountingConfig,
|
|
LetterCountingCurriculum,
|
|
LetterCountingDataset,
|
|
)
|
|
|
|
|
|
def test_letter_counting_config_validation():
|
|
"""Test that invalid configs raise appropriate errors"""
|
|
with pytest.raises(AssertionError):
|
|
config = LetterCountingConfig(min_words=0)
|
|
config.validate()
|
|
|
|
with pytest.raises(AssertionError):
|
|
config = LetterCountingConfig(min_words=10, max_words=5)
|
|
config.validate()
|
|
|
|
|
|
def test_letter_counting_dataset_deterministic():
|
|
"""Test that dataset generates same items with same seed"""
|
|
config = LetterCountingConfig(seed=42, size=10)
|
|
dataset1 = LetterCountingDataset(config)
|
|
dataset2 = LetterCountingDataset(config)
|
|
|
|
for i in range(len(dataset1)):
|
|
assert dataset1[i] == dataset2[i]
|
|
|
|
|
|
def test_letter_counting_dataset_items():
|
|
"""Test basic properties of generated items"""
|
|
config = LetterCountingConfig(min_words=3, max_words=6, size=10, seed=42)
|
|
dataset = LetterCountingDataset(config)
|
|
|
|
for i in range(len(dataset)):
|
|
item = dataset[i]
|
|
# Check item structure
|
|
assert isinstance(item, dict)
|
|
assert "question" in item
|
|
assert "answer" in item
|
|
assert "metadata" in item
|
|
|
|
# Check metadata
|
|
assert "span_length" in item["metadata"]
|
|
assert "target_letter" in item["metadata"]
|
|
assert "span" in item["metadata"]
|
|
|
|
# Verify span length constraints
|
|
span = item["metadata"]["span"]
|
|
assert len(span) >= config.min_words
|
|
assert len(span) <= config.max_words
|
|
|
|
# Verify letter counting
|
|
target_letter = item["metadata"]["target_letter"]
|
|
count = sum(word.lower().count(target_letter) for word in span)
|
|
assert str(count) == item["answer"]
|
|
|
|
|
|
def test_letter_counting_dataset_iteration():
|
|
"""Test that iteration respects dataset size"""
|
|
config = LetterCountingConfig(size=5, seed=42)
|
|
dataset = LetterCountingDataset(config)
|
|
|
|
items = list(dataset)
|
|
assert len(items) == config.size
|
|
|
|
# Test multiple iterations yield same items
|
|
assert items == list(dataset)
|
|
|
|
|
|
def test_letter_counting_text_preprocessing():
|
|
"""Test that text preprocessing handles edge cases"""
|
|
config = LetterCountingConfig(size=1, seed=42)
|
|
dataset = LetterCountingDataset(config)
|
|
|
|
# Verify words were extracted from text
|
|
assert len(dataset.words) > 0
|
|
# Verify words contain only word characters
|
|
assert all(word.isalnum() for word in dataset.words)
|
|
|
|
|
|
def test_letter_counting_curriculum():
|
|
curriculum = LetterCountingCurriculum()
|
|
|
|
base_value = {"size": 150, "seed": 1}
|
|
|
|
base_cfg: LetterCountingConfig = curriculum.generate_configuration(base_value)
|
|
assert base_cfg.seed == 1
|
|
assert base_cfg.size == 150
|
|
assert base_cfg.min_words == 5 and base_cfg.max_words == 7
|
|
|
|
# test incrementing attribute levels
|
|
curriculum.increment_attr_level("words")
|
|
increased_cfg = curriculum.generate_configuration(base_value)
|
|
assert increased_cfg.min_words == 5 and increased_cfg.max_words == 9
|
|
|
|
# test decrementing attribute level for words again
|
|
curriculum.decrement_attr_level("words")
|
|
partially_decreased_cfg = curriculum.generate_configuration(base_value)
|
|
assert partially_decreased_cfg.min_words == 5 and partially_decreased_cfg.max_words == 7
|