import math

import pytest
from unittest.mock import MagicMock, patch, mock_open
from AAA_VocabularyLevelEstimator import VocabularyLevelEstimator

# 模拟加载记录函数返回的数据
mock_words_and_tests = {'word1': ['CET4'], 'word2': ['OXFORD3000']}
mock_word_vectors_path = './wiki-news-300d-1M.vec'


@pytest.fixture
def estimator():
    with patch('gensim.models.KeyedVectors.load_word2vec_format', return_value=MagicMock()):
        return VocabularyLevelEstimator(word_vectors_path=mock_word_vectors_path)


@pytest.fixture
def estimator_with_failed_load():
    with patch('gensim.models.KeyedVectors.load_word2vec_format', side_effect=Exception("Failed to load word vectors")):
        return VocabularyLevelEstimator(word_vectors_path=mock_word_vectors_path)


def test_estimate_text_difficulty_normal(estimator):
    text = "This is a simple sentence with common words."
    difficulty = estimator.estimate_text_difficulty(text)
    assert isinstance(difficulty, float) and not math.isnan(difficulty)


def test_estimate_user_vocabulary_level_normal(estimator):
    user_unknown_words = ['abandon', 'pursuit', 'knowledge']
    level = estimator.estimate_user_vocabulary_level(user_unknown_words)
    assert isinstance(level, float) and not math.isnan(level)


def test_estimate_text_difficulty_empty_string(estimator):
    text = ""
    difficulty = estimator.estimate_text_difficulty(text)
    assert difficulty == 0.0


def test_estimate_user_vocabulary_level_empty_list(estimator):
    user_unknown_words = []
    level = estimator.estimate_user_vocabulary_level(user_unknown_words)
    assert level == 0.0


def test_estimate_text_difficulty_with_stopwords_only(estimator):
    text = "the and of to in"
    difficulty = estimator.estimate_text_difficulty(text)
    assert difficulty == 0.0


def test_estimate_word_difficulty_cache(estimator):
    word = "example"
    first_call = estimator.estimate_word_difficulty(word)
    second_call = estimator.estimate_word_difficulty(word)
    assert first_call == second_call


def test_estimate_text_difficulty_non_english_characters(estimator):
    text = "你好，世界！"
    difficulty = estimator.estimate_text_difficulty(text)
    assert difficulty == 0.0


def test_estimate_user_vocabulary_level_non_english_characters(estimator):
    user_unknown_words = ['你好', '世界']
    level = estimator.estimate_user_vocabulary_level(user_unknown_words)
    assert level == 0.0


def test_estimate_text_difficulty_exception_handling(estimator):
    original_method = estimator.estimate_word_difficulty
    estimator.estimate_word_difficulty = MagicMock(side_effect=Exception("Simulated error"))

    text = "This should trigger exception handling."
    difficulty = estimator.estimate_text_difficulty(text)
    assert isinstance(difficulty, float) and not math.isnan(difficulty)

    estimator.estimate_word_difficulty = original_method


def test_estimator_initialization_with_failed_word_vector_load(estimator_with_failed_load):
    """确保即使加载词向量失败，类也能正常初始化"""
    assert estimator_with_failed_load.word_vectors is None


if __name__ == '__main__':
    pytest.main()