def test_probability_language_detector_detect_language_ideal(self): unknown_file = open('lab_3/unknown_Arthur_Conan_Doyle.txt', encoding='utf-8') german_file = open('lab_3/Thomas_Mann.txt', encoding='utf-8') english_file = open('lab_3/Frank_Baum.txt', encoding='utf-8') text_unk = tokenize_by_sentence(unknown_file.read()) text_ger = tokenize_by_sentence(german_file.read()) text_eng = tokenize_by_sentence(english_file.read()) english_file.close() german_file.close() unknown_file.close() letter_storage = LetterStorage() letter_storage.update(text_eng) letter_storage.update(text_ger) letter_storage.update(text_unk) eng_encoded = encode_corpus(letter_storage, text_eng) unk_encoded = encode_corpus(letter_storage, text_unk) ger_encoded = encode_corpus(letter_storage, text_ger) language_detector = ProbabilityLanguageDetector((3, 4, 5), 1000) language_detector.new_language(eng_encoded, 'english') language_detector.new_language(ger_encoded, 'german') actual = language_detector.detect_language(unk_encoded) self.assertTrue(actual['german'] > actual['english'])
def test_probability_language_detector_calls_required_method(self, mock): unknown_file = open('lab_3/unknown_Arthur_Conan_Doyle.txt', encoding='utf-8') german_file = open('lab_3/Thomas_Mann.txt', encoding='utf-8') english_file = open('lab_3/Frank_Baum.txt', encoding='utf-8') text_unk = tokenize_by_sentence(unknown_file.read()) text_ger = tokenize_by_sentence(german_file.read()) text_eng = tokenize_by_sentence(english_file.read()) english_file.close() german_file.close() unknown_file.close() letter_storage = LetterStorage() letter_storage.update(text_eng) letter_storage.update(text_ger) letter_storage.update(text_unk) eng_encoded = encode_corpus(letter_storage, text_eng) unk_encoded = encode_corpus(letter_storage, text_unk) ger_encoded = encode_corpus(letter_storage, text_ger) language_detector = ProbabilityLanguageDetector((3, 4, 5), 1000) language_detector.new_language(eng_encoded, 'english') language_detector.new_language(ger_encoded, 'german') ngram_unknown = NGramTrie(4) ngram_unknown.fill_n_grams(unk_encoded) language_detector.detect_language(ngram_unknown.n_grams) self.assertTrue(mock.called)
def test_new_language_storage_already_created(self): letter_storage = LetterStorage() language_detector = LanguageDetector((3, ), 10) file = open('lab_3/Thomas_Mann.txt', 'r', encoding='utf-8') file_unknown = open('lab_3/unknown_Arthur_Conan_Doyle.txt', 'r', encoding='utf-8') text = tokenize_by_sentence(file.read()) text_unknown = tokenize_by_sentence(file_unknown.read()) letter_storage.update(text) letter_storage.update(text_unknown) encoded_text = encode_corpus(letter_storage, text) encoded_unknown_text = encode_corpus(letter_storage, text_unknown) file.close() file_unknown.close() language_detector.new_language(encoded_text, 'german') language_detector.new_language(encoded_unknown_text, 'english') self.assertTrue(language_detector.n_gram_storages['german']) self.assertTrue(language_detector.n_gram_storages['english']) self.assertEqual(type(language_detector.n_gram_storages['german'][3]), NGramTrie) self.assertEqual(type(language_detector.n_gram_storages['english'][3]), NGramTrie)
def test_detect_language_uses_several_ngrams(self): letter_storage = LetterStorage() language_detector = LanguageDetector((2, 3), 100) file_first = open('lab_3/Frank_Baum.txt', 'r', encoding='utf-8') file_second = open('lab_3/Thomas_Mann.txt', 'r', encoding='utf-8') file_third = open('lab_3/unknown_Arthur_Conan_Doyle.txt', 'r', encoding='utf-8') text_english = tokenize_by_sentence(file_first.read()) text_german = tokenize_by_sentence(file_second.read()) text_unknown = tokenize_by_sentence(file_third.read()) letter_storage.update(text_english) letter_storage.update(text_german) letter_storage.update(text_unknown) encoded_english = encode_corpus(letter_storage, text_english) encoded_german = encode_corpus(letter_storage, text_german) encoded_unknown = encode_corpus(letter_storage, text_unknown) file_first.close() file_second.close() file_third.close() language_detector.new_language(encoded_english, 'english') language_detector.new_language(encoded_german, 'german') actual = language_detector.detect_language(encoded_unknown) self.assertTrue(actual['german'] > actual['english'])
def test_tokenize_by_sentence_inappropriate_sentence(self): """ Tests that tokenize_by_sentence function can handle inappropriate sentence input """ text = '$#&*@#$*#@)' expected = () actual = tokenize_by_sentence(text) self.assertEqual(expected, actual)
def test_tokenize_by_sentence_empty_sentence(self): """ Tests that tokenize_by_sentence function can handle empty sentence input """ text = '' expected = () actual = tokenize_by_sentence(text) self.assertEqual(expected, actual)
def test_tokenize_by_sentence_incorrect_input(self): """ Tests that tokenize_by_sentence function can handle incorrect input cases """ bad_inputs = [[], {}, (), None, 9, 9.34, True] expected = () for bad_input in bad_inputs: actual = tokenize_by_sentence(bad_input) self.assertEqual(expected, actual)
def test_tokenize_by_sentence_ideal(self): """ Tests that tokenize_by_sentence function can handle ideal two sentence input """ text = 'She is happy. He is happy.' expected = ((('_', 's', 'h', 'e', '_'), ('_', 'i', 's', '_'), ('_', 'h', 'a', 'p', 'p', 'y', '_')), (('_', 'h', 'e', '_'), ('_', 'i', 's', '_'), ('_', 'h', 'a', 'p', 'p', 'y', '_'))) actual = tokenize_by_sentence(text) self.assertEqual(expected, actual)
def test_detect_language_calls_required_method(self, mock): letter_storage = LetterStorage() language_detector = LanguageDetector((3, ), 10) text_to_detect = (((1, 2, 3), ), ) file = open('lab_3/Frank_Baum.txt', 'r', encoding='utf-8') text = tokenize_by_sentence(file.read()) letter_storage.update(text) encoded_text = encode_corpus(letter_storage, text) file.close() language_detector.new_language(encoded_text, 'english') language_detector.detect_language(text_to_detect) self.assertTrue(mock.called)
def test_probability_language_detector_calculate_probability_ideal(self): print('launching test') english_file = open('lab_3/Frank_Baum.txt', encoding='utf-8') german_file = open('lab_3/Thomas_Mann.txt', encoding='utf-8') unknown_file = open('lab_3/unknown_Arthur_Conan_Doyle.txt', encoding='utf-8') english_text = tokenize_by_sentence(english_file.read()) german_text = tokenize_by_sentence(german_file.read()) unknown_text = tokenize_by_sentence(unknown_file.read()) english_file.close() german_file.close() unknown_file.close() letter_storage = LetterStorage() letter_storage.update(english_text) letter_storage.update(german_text) letter_storage.update(unknown_text) english_encoded = encode_corpus(letter_storage, english_text) german_encoded = encode_corpus(letter_storage, german_text) unknown_encoded = encode_corpus(letter_storage, unknown_text) language_detector = ProbabilityLanguageDetector((3,), 1000) language_detector.new_language(english_encoded, 'english') language_detector.new_language(german_encoded, 'german') n3_gram_trie_english = language_detector.n_gram_storages['english'][3] n3_gram_trie_german = language_detector.n_gram_storages['german'][3] n3_gram_unknown = NGramTrie(3) n3_gram_unknown.fill_n_grams(unknown_encoded) english_prob = language_detector._calculate_sentence_probability(n3_gram_trie_english, n3_gram_unknown.n_grams) german_prob = language_detector._calculate_sentence_probability(n3_gram_trie_german, n3_gram_unknown.n_grams) print(f'English_sentence_prob: {english_prob}') print(f'Deutsch_sentence_prob: {german_prob}') self.assertTrue(english_prob > german_prob)
def test_probability_language_detector_several_ngrams_case(self): language_detector = ProbabilityLanguageDetector((3, 5), 1000) english_file = open('lab_3/Frank_Baum.txt', encoding='utf-8') german_file = open('lab_3/Thomas_Mann.txt', encoding='utf-8') unknown_file = open('lab_3/unknown_Arthur_Conan_Doyle.txt', encoding='utf-8') eng_text = tokenize_by_sentence(english_file.read()) ger_text = tokenize_by_sentence(german_file.read()) unk_text = tokenize_by_sentence(unknown_file.read()) english_file.close() german_file.close() unknown_file.close() letter_storage = LetterStorage() letter_storage.update(eng_text) letter_storage.update(ger_text) letter_storage.update(unk_text) english_encoded = encode_corpus(letter_storage, eng_text) german_encoded = encode_corpus(letter_storage, ger_text) unknown_encoded = encode_corpus(letter_storage, unk_text) language_detector.new_language(english_encoded, 'english') language_detector.new_language(german_encoded, 'german') eng_prob = language_detector.n_gram_storages['english'][5] ger_prob = language_detector.n_gram_storages['german'][5] ngram_trie = NGramTrie(5) ngram_trie.fill_n_grams(unknown_encoded) eng = language_detector._calculate_sentence_probability( eng_prob, ngram_trie.n_grams) ger = language_detector._calculate_sentence_probability( ger_prob, ngram_trie.n_grams) self.assertTrue(ger > eng)
def test_new_language_creates_several_ngrams(self): letter_storage = LetterStorage() language_detector = LanguageDetector((2, 3), 10) file = open('lab_3/Frank_Baum.txt', 'r', encoding='utf-8') text = tokenize_by_sentence(file.read()) letter_storage.update(text) encoded_text = encode_corpus(letter_storage, text) file.close() language_detector.new_language(encoded_text, 'english') self.assertTrue(language_detector.n_gram_storages['english'][2]) self.assertTrue(language_detector.n_gram_storages['english'][3])
def test_new_language_add_existing_language(self): letter_storage = LetterStorage() language_detector = LanguageDetector((3, ), 10) file = open('lab_3/Frank_Baum.txt', 'r', encoding='utf-8') text = tokenize_by_sentence(file.read()) letter_storage.update(text) encoded_text = encode_corpus(letter_storage, text) file.close() expected = 0 language_detector.new_language(encoded_text, 'german') actual = language_detector.new_language(encoded_text, 'german') self.assertEqual(expected, actual)
def test_tokenize_by_sentence_dirty_text(self): """ Tests that tokenize_by_sentence function can handle text filled with inappropriate characters """ text = 'The first% sentence><. The sec&*ond sent@ence #.' expected = ((('_', 't', 'h', 'e', '_'), ('_', 'f', 'i', 'r', 's', 't', '_'), ('_', 's', 'e', 'n', 't', 'e', 'n', 'c', 'e', '_')), (('_', 't', 'h', 'e', '_'), ('_', 's', 'e', 'c', 'o', 'n', 'd', '_'), ('_', 's', 'e', 'n', 't', 'e', 'n', 'c', 'e', '_'))) actual = tokenize_by_sentence(text) self.assertEqual(expected, actual)
def test_tokenize_by_sentence_punctuation_marks(self): """ Tests that tokenize_by_sentence function can process and ignore different punctuation marks """ text = 'The, first sentence - nice. The second sentence: bad!' expected = ((('_', 't', 'h', 'e', '_'), ('_', 'f', 'i', 'r', 's', 't', '_'), ('_', 's', 'e', 'n', 't', 'e', 'n', 'c', 'e', '_'), ('_', 'n', 'i', 'c', 'e', '_')), (('_', 't', 'h', 'e', '_'), ('_', 's', 'e', 'c', 'o', 'n', 'd', '_'), ('_', 's', 'e', 'n', 't', 'e', 'n', 'c', 'e', '_'), ('_', 'b', 'a', 'd', '_'))) actual = tokenize_by_sentence(text) self.assertEqual(expected, actual)
def test_tokenize_by_sentence_complex(self): """ Tests that tokenize_by_sentence function can handle complex split case """ text = 'Mar#y wa$nted, to swim. However, she was afraid of sharks.' expected = ((('_', 'm', 'a', 'r', 'y', '_'), ('_', 'w', 'a', 'n', 't', 'e', 'd', '_'), ('_', 't', 'o', '_'), ('_', 's', 'w', 'i', 'm', '_')), (('_', 'h', 'o', 'w', 'e', 'v', 'e', 'r', '_'), ('_', 's', 'h', 'e', '_'), ('_', 'w', 'a', 's', '_'), ('_', 'a', 'f', 'r', 'a', 'i', 'd', '_'), ('_', 'o', 'f', '_'), ('_', 's', 'h', 'a', 'r', 'k', 's', '_'))) actual = tokenize_by_sentence(text) self.assertEqual(expected, actual)
from lab_3.main import encode_corpus from lab_3.main import LetterStorage if __name__ == '__main__': # here goes your function calls letter_storage = LetterStorage() language_detector = LanguageDetector((3, ), 100) file_first = open('lab_3/Frank_Baum.txt', 'r', encoding='utf-8') file_second = open('lab_3/Thomas_Mann.txt', 'r', encoding='utf-8') file_third = open('lab_3/unknown_Arthur_Conan_Doyle.txt', 'r', encoding='utf-8') text_english = tokenize_by_sentence(file_first.read()) text_german = tokenize_by_sentence(file_second.read()) text_unknown = tokenize_by_sentence(file_third.read()) letter_storage.update(text_english) letter_storage.update(text_german) letter_storage.update(text_unknown) encoded_english = encode_corpus(letter_storage, text_english) encoded_german = encode_corpus(letter_storage, text_german) encoded_unknown = encode_corpus(letter_storage, text_unknown) file_first.close() file_second.close() file_third.close() language_detector.new_language(encoded_english, 'english') language_detector.new_language(encoded_german, 'german')
Language detector implementation starter """ from lab_3.main import tokenize_by_sentence from lab_3.main import encode_corpus from lab_3.main import NGramTrie from lab_3.main import LetterStorage from lab_3.main import LanguageDetector if __name__ == '__main__': unknown_file = open('lab_3/unknown_Arthur_Conan_Doyle.txt', encoding='utf-8') german_file = open('lab_3/Thomas_Mann.txt', encoding='utf-8') english_file = open('lab_3/Frank_Baum.txt', encoding='utf-8') text_unk = tokenize_by_sentence(unknown_file.read()) text_ger = tokenize_by_sentence(german_file.read()) text_eng = tokenize_by_sentence(english_file.read()) english_file.close() german_file.close() unknown_file.close() letter_storage = LetterStorage() letter_storage.update(text_eng) letter_storage.update(text_ger) letter_storage.update(text_unk) eng_encoded = encode_corpus(letter_storage, text_eng) unk_encoded = encode_corpus(letter_storage, text_unk) ger_encoded = encode_corpus(letter_storage, text_ger)