def generate_keras_extractor(documents):
cleaner = clean.DocumentCleaner()
documents = cleaner.clean(documents, True)
extractor = keras_extractor.KerasLSTMFeatureExtractor()
extractor.build_feature_set(documents)
return extractor
def generate_bag_of_ngrams_extractor(documents, nfeatures, ngrams):
cleaner = clean.DocumentCleaner()
documents = cleaner.clean(documents, True)
extractor = bag_of_ngrams_extractor.BagOfNGramsFeatureExtractor()
extractor.build_feature_set(documents, nfeatures, ngrams)
return extractor
def generate_nb_input(documents, labels, label_set):
SHOULD_ADD_NEGATIONS = True
cleaner = clean.DocumentCleaner()
documents = cleaner.clean(documents, SHOULD_ADD_NEGATIONS)
nb_input = partition_documents_by_class(documents, labels, label_set)
return nb_input
def generate_glove_extractor(documents, nfeatures):
SHOULD_ADD_NEGATIONS = False
cleaner = clean.DocumentCleaner()
documents = cleaner.clean(documents, SHOULD_ADD_NEGATIONS)
extractor = glove_extractor.GloveFeatureExtractor()
extractor.build_feature_set(documents, nfeatures)
return extractor
def generate_input(documents, extractor, SHOULD_ADD_NEGATIONS=True):
cleaner = clean.DocumentCleaner()
documents = cleaner.clean(documents, SHOULD_ADD_NEGATIONS)
input = extractor.extract_features(documents)
return input