CountVectorizer: Vocabulary wasn't fitted

Question

I instantiated a sklearn.feature_extraction.text.CountVectorizer object by passing a vocabulary through the vocabulary argument, but I get a sklearn.utils.val

Answer 1

For some reason, even though you passed vocabulary=vocabulary_to_load as argument for sklearn.feature_extraction.text.CountVectorizer(), you still need to call loaded_vectorizer._validate_vocabulary() before being able to call loaded_vectorizer.get_feature_names().

In your example, you should therefore do the following when creating an CountVectorizer object with your vocabulary:

vocabulary_to_load = pickle.load(open(dictionary_filepath, 'r'))
loaded_vectorizer = sklearn.feature_extraction.text.CountVectorizer(ngram_range=(ngram_size,
                                        ngram_size), min_df=1, vocabulary=vocabulary_to_load)
loaded_vectorizer._validate_vocabulary()
print('loaded_vectorizer.get_feature_names(): {0}'.
  format(loaded_vectorizer.get_feature_names()))