Python word2vec.load方法代码示例

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def transfer(model_path, embedding_size):
    start_time = time.time()
    model = word2vec.load(model_path)
    word2id_dic = {}
    init_0 = [0.0 for i in range(embedding_size)]
    id2vec_dic = [init_0]
    for i in range(len(model.vocab)):
        id = i + 1
        word2id_dic[model.vocab[i]] = id
        id2vec_dic.append(model[model.vocab[i]].tolist())
    end_time = time.time()
    print('词转id，id转向量完成')
    print(end_time - start_time)
    return word2id_dic, id2vec_dic

# 存入json文件 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def main(em_file, em_result):
    '''
    embedding ->numpy
    '''
    em = word2vec.load(em_file)
    vec = (em.vectors)
    word2id = em.vocab_hash
    # d = dict(vector = vec, word2id = word2id)
    # t.save(d,em_result)
    np.savez_compressed(em_result,vector=vec,word2id=word2id) 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def emb2npz(emb_file_path, emb_dict_path):
    """将txt格式的embedding转为字典格式, 并将<PAD>和<UNK>加入"""
    emb = word2vec.load(emb_file_path)
    vec = emb.vectors
    word2id = emb.vocab_hash
    word2id['<PAD>'] = len(word2id)
    pad_row = [0] * vec.shape[1]
    vec = np.row_stack((vec, pad_row))
    np.savez_compressed(emb_dict_path, vec=vec, word2id=word2id)
    print('word size: {}'.format(len(word2id)))
    print('emb shape: {}'.format(vec.shape)) 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def data2npz(src_path, dst_path):
    """src_path txt: label+\t+title+\t+content
    如：40,6  w6061,w26959,w109   w23255,w728,w12768,w58588,w11,w1442,w855,w36791"""

    data = np.load(conf.emb_path)
    word2id = data['word2id'].item()
    del data

    labels = []
    titles = []
    contents = []
    with open(src_path, 'r', encoding='utf-8') as f:
        for i, line in enumerate(f):
            label, title, content = line.replace('\n', '').split('\t')

            label = [int(lab) for lab in label.split(',')]
            label_mat = np.zeros(conf.n_classes, dtype='int32')
            label_mat[label] = 1
            labels.append(label_mat)

            # word2id
            title = [word2id[word if word in word2id else '</s>'] for word in title.split(',') if word.rstrip()]
            content = [word2id[word if word in word2id else '</s>'] for word in content.split(',') if word.rstrip()]
            # padding
            titles.append(padding(title, conf.title_seq_len, pad=word2id['<PAD>']))
            contents.append(padding(content, conf.content_seq_len, pad=word2id['<PAD>']))
    print('data size: {}'.format(len(labels)))
    np.savez_compressed(dst_path, label=labels, title=titles, content=contents) 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def load_word2vec(self, word2vec_file):
        model = word2vec.load(word2vec_file)

        embedding_matrix = []

        for word in self.vocab:
            try:
                embedding_matrix.append(model[word])
            except KeyError:
                # print(word)
                embedding_matrix.append(model['the'])

        embedding_matrix = np.array(embedding_matrix)

        return embedding_matrix 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def transfer(model_path, embedding_size):
    start_time = time.time()
    model = word2vec.load(model_path)
    word2id_dic = {}
    init_0 = [0.0 for i in range(embedding_size)]
    id2vec_dic = [init_0]
    for i in range(len(model.vocab)):
        id = i + 1
        word2id_dic[model.vocab[i]] = id
        id2vec_dic.append(model[model.vocab[i]].tolist())
    end_time = time.time()
    print('词转id，id转向量完成')
    print(end_time - start_time)
    return word2id_dic, id2vec_dic 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def __init__(self,file_path):
        # w2v_file = os.path.join(base_path, "vectors_poem.bin")
        self.model = word2vec.load(file_path)
        self.add_word('<unknown>')
        self.add_word('<pad>')
        # self.vocab_size = len(self.model.vocab) 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def __init__(self, vector_path):
        self.vector_path = vector_path
        if vector_path.endswith('.csv'):
            self.vectors = pd.read_csv(vector_path).set_index('pfam_id')
        elif vector_path.endswith('.tsv'):
            self.vectors = pd.read_csv(vector_path, sep='\t').set_index('pfam_id')
        elif vector_path.endswith('.pkl') or vector_path.endswith('.pickle'):
            self.vectors = pd.read_pickle(vector_path)
        elif vector_path.endswith('.bin'):
            import word2vec
            model = word2vec.load(vector_path, kind='bin')
            self.vectors = pd.DataFrame(model.vectors, index=model.vocab)
        else:
            raise ValueError("File type {} not supported for Pfam2Vec, use .csv, .pkl, .pickle or .bin".format(vector_path))

        vectors_min = self.vectors.min()
        too_low_idx = vectors_min < -1
        cols_too_low = list(self.vectors.columns[too_low_idx])
        if cols_too_low:
            raise ValueError('Pfam2vec vectors should be >= -1, got {} in {}'.format(list(vectors_min[too_low_idx]), cols_too_low))

        vectors_max = self.vectors.max()
        too_high_idx = vectors_max > 1
        cols_too_high = list(self.vectors.columns[too_high_idx])
        if cols_too_high:
            raise ValueError('Pfam2vec vectors should be <= 1, got {} in {}'.format(list(vectors_max[too_high_idx]), cols_too_high)) 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def create_voabulary(simple=None,word2vec_model_path='../zhihu-word2vec-title-desc.bin-100',name_scope=''): #zhihu-word2vec-multilabel.bin-100
    cache_path ='../cache_vocabulary_label_pik/'+ name_scope + "_word_voabulary.pik"
    print("cache_path:",cache_path,"file_exists:",os.path.exists(cache_path))
    if os.path.exists(cache_path):#如果缓存文件存在，则直接读取
        with open(cache_path, 'r') as data_f:
            vocabulary_word2index, vocabulary_index2word=pickle.load(data_f)
            return vocabulary_word2index, vocabulary_index2word
    else:
        vocabulary_word2index={}
        vocabulary_index2word={}
        if simple is not None:
            word2vec_model_path='../zhihu-word2vec.bin-100'
        print("create vocabulary. word2vec_model_path:",word2vec_model_path)
        model=word2vec.load(word2vec_model_path,kind='bin')
        vocabulary_word2index['PAD_ID']=0
        vocabulary_index2word[0]='PAD_ID'
        special_index=0
        if 'biLstmTextRelation' in name_scope:
            vocabulary_word2index['EOS']=1 # a special token for biLstTextRelation model. which is used between two sentences.
            vocabulary_index2word[1]='EOS'
            special_index=1
        for i,vocab in enumerate(model.vocab):
            vocabulary_word2index[vocab]=i+1+special_index
            vocabulary_index2word[i+1+special_index]=vocab

        #save to file system if vocabulary of words is not exists.
        if not os.path.exists(cache_path): #如果不存在写到缓存文件中
            with open(cache_path, 'a') as data_f:
                pickle.dump((vocabulary_word2index,vocabulary_index2word), data_f)
    return vocabulary_word2index,vocabulary_index2word

# create vocabulary of lables. label is sorted. 1 is high frequency, 2 is low frequency. 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def create_voabulary_labelO():
    model = word2vec.load('zhihu-word2vec-multilabel.bin-100', kind='bin') #zhihu-word2vec.bin-100
    count=0
    vocabulary_word2index_label={}
    vocabulary_index2word_label={}
    label_unique={}
    for i,vocab in enumerate(model.vocab):
        if '__label__' in vocab:  #'__label__-2051131023989903826
            label=vocab[vocab.index('__label__')+len('__label__'):]
            if label_unique.get(label,None) is None: #不曾出现过的话，保持到字典中
                vocabulary_word2index_label[label]=count
                vocabulary_index2word_label[count]=label #ADD
                count=count+1
                label_unique[label]=label
    return vocabulary_word2index_label,vocabulary_index2word_label 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def load_data(vocabulary_word2index,vocabulary_word2index_label,valid_portion=0.05,max_training_data=1000000,training_data_path='train-zhihu4-only-title-all.txt'):  # n_words=100000,
    """
    input: a file path
    :return: train, test, valid. where train=(trainX, trainY). where
                trainX: is a list of list.each list representation a sentence.trainY: is a list of label. each label is a number
    """
    # 1.load a zhihu data from file
    # example:"w305 w6651 w3974 w1005 w54 w109 w110 w3974 w29 w25 w1513 w3645 w6 w111 __label__-400525901828896492"
    print("load_data.started...")
    zhihu_f = codecs.open(training_data_path, 'r', 'utf8') #-zhihu4-only-title.txt
    lines = zhihu_f.readlines()
    # 2.transform X as indices
    # 3.transform  y as scalar
    X = []
    Y = []
    for i, line in enumerate(lines):
        x, y = line.split('__label__') #x='w17314 w5521 w7729 w767 w10147 w111'
        y=y.replace('\n','')
        x = x.replace("\t",' EOS ').strip()
        if i<5:
            print("x0:",x) #get raw x
        #x_=process_one_sentence_to_get_ui_bi_tri_gram(x)
        #if i<5:
        #    print("x1:",x_) #
        x=x.split(" ")
        x = [vocabulary_word2index.get(e,0) for e in x] #if can't find the word, set the index as '0'.(equal to PAD_ID = 0)
        if i<5:
            print("x1:",x) #word to index
        y = vocabulary_word2index_label[y] #np.abs(hash(y))
        X.append(x)
        Y.append(y)
    # 4.split to train,test and valid data
    number_examples = len(X)
    print("number_examples:",number_examples) #
    train = (X[0:int((1 - valid_portion) * number_examples)], Y[0:int((1 - valid_portion) * number_examples)])
    test = (X[int((1 - valid_portion) * number_examples) + 1:], Y[int((1 - valid_portion) * number_examples) + 1:])
    # 5.return
    print("load_data.ended...")
    return train, test, test

 # 将一句话转化为(uigram,bigram,trigram)后的字符串 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def assign_pretrained_word_embedding(sess,vocabulary_index2word,vocab_size,model,word2vec_model_path=None):
    print("using pre-trained word emebedding.started.word2vec_model_path:",word2vec_model_path)
    # word2vecc=word2vec.load('word_embedding.txt') #load vocab-vector fiel.word2vecc['w91874']
    word2vec_model = word2vec.load(word2vec_model_path, kind='bin')
    word2vec_dict = {}
    for word, vector in zip(word2vec_model.vocab, word2vec_model.vectors):
        word2vec_dict[word] = vector
    word_embedding_2dlist = [[]] * vocab_size  # create an empty word_embedding list.
    word_embedding_2dlist[0] = np.zeros(FLAGS.embed_size)  # assign empty for first word:'PAD'
    bound = np.sqrt(6.0) / np.sqrt(vocab_size)  # bound for random variables.
    count_exist = 0;
    count_not_exist = 0
    for i in range(1, vocab_size):  # loop each word
        word = vocabulary_index2word[i]  # get a word
        embedding = None
        try:
            embedding = word2vec_dict[word]  # try to get vector:it is an array.
        except Exception:
            embedding = None
        if embedding is not None:  # the 'word' exist a embedding
            word_embedding_2dlist[i] = embedding;
            count_exist = count_exist + 1  # assign array to this word.
        else:  # no embedding for this word
            word_embedding_2dlist[i] = np.random.uniform(-bound, bound, FLAGS.embed_size);
            count_not_exist = count_not_exist + 1  # init a random value for the word.
    word_embedding_final = np.array(word_embedding_2dlist)  # covert to 2d array.
    word_embedding = tf.constant(word_embedding_final, dtype=tf.float32)  # convert to tensor
    t_assign_embedding = tf.assign(model.Embedding,word_embedding)  # assign this value to our embedding variables of our model.
    sess.run(t_assign_embedding);
    print("word. exists embedding:", count_exist, " ;word not exist embedding:", count_not_exist)
    print("using pre-trained word emebedding.ended...")

# 在验证集上做验证，报告损失、精确度 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def create_voabulary(simple=None,word2vec_model_path='zhihu-word2vec-title-desc.bin-100',name_scope=''):
    cache_path ='cache_vocabulary_label_pik/'+ name_scope + "_word_voabulary.pik"
    print("cache_path:",cache_path,"file_exists:",os.path.exists(cache_path))
    if os.path.exists(cache_path):#if exists, load it; otherwise create it.
        with open(cache_path, 'r') as data_f:
            vocabulary_word2index, vocabulary_index2word=pickle.load(data_f)
            return vocabulary_word2index, vocabulary_index2word
    else:
        vocabulary_word2index={}
        vocabulary_index2word={}
        if simple is not None:
            word2vec_model_path='zhihu-word2vec.bin-100'
        print("create vocabulary. word2vec_model_path:",word2vec_model_path)
        model=word2vec.load(word2vec_model_path,kind='bin')
        vocabulary_word2index['PAD_ID']=0
        vocabulary_index2word[0]='PAD_ID'
        special_index=0
        if 'biLstmTextRelation' in name_scope:
            vocabulary_word2index['EOS']=1 # a special token for biLstTextRelation model. which is used between two sentences.
            vocabulary_index2word[1]='EOS'
            special_index=1
        for i,vocab in enumerate(model.vocab):
            vocabulary_word2index[vocab]=i+1+special_index
            vocabulary_index2word[i+1+special_index]=vocab

        #save to file system if vocabulary of words is not exists.
        if not os.path.exists(cache_path): #如果不存在写到缓存文件中
            with open(cache_path, 'a') as data_f:
                pickle.dump((vocabulary_word2index,vocabulary_index2word), data_f)
    return vocabulary_word2index,vocabulary_index2word

# create vocabulary of lables. label is sorted. 1 is high frequency, 2 is low frequency. 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def process_one_sentence_to_get_ui_bi_tri_gram(sentence,n_gram=3):
    """
    :param sentence: string. example:'w17314 w5521 w7729 w767 w10147 w111'
    :param n_gram:
    :return:string. example:'w17314 w17314w5521 w17314w5521w7729 w5521 w5521w7729 w5521w7729w767 w7729 w7729w767 w7729w767w10147 w767 w767w10147 w767w10147w111 w10147 w10147w111 w111'
    """
    result=[]
    word_list=sentence.split(" ") #[sentence[i] for i in range(len(sentence))]
    unigram='';bigram='';trigram='';fourgram=''
    length_sentence=len(word_list)
    for i,word in enumerate(word_list):
        unigram=word                           #ui-gram
        word_i=unigram
        if n_gram>=2 and i+2<=length_sentence: #bi-gram
            bigram="".join(word_list[i:i+2])
            word_i=word_i+' '+bigram
        if n_gram>=3 and i+3<=length_sentence: #tri-gram
            trigram="".join(word_list[i:i+3])
            word_i = word_i + ' ' + trigram
        if n_gram>=4 and i+4<=length_sentence: #four-gram
            fourgram="".join(word_list[i:i+4])
            word_i = word_i + ' ' + fourgram
        if n_gram>=5 and i+5<=length_sentence: #five-gram
            fivegram="".join(word_list[i:i+5])
            word_i = word_i + ' ' + fivegram
        result.append(word_i)
    result=" ".join(result)
    return result

# 加载数据，标签包含多个label：load data with multi-labels 

# 需要导入模块: import word2vec [as 别名]
# 或者: from word2vec import load [as 别名]
def assign_pretrained_word_embedding(sess,vocabulary_index2word,vocab_size,textCNN,word2vec_model_path=None):
    print("using pre-trained word emebedding.started.word2vec_model_path:",word2vec_model_path)
    # word2vecc=word2vec.load('word_embedding.txt') #load vocab-vector fiel.word2vecc['w91874']
    word2vec_model = word2vec.load(word2vec_model_path, kind='bin')
    word2vec_dict = {}
    for word, vector in zip(word2vec_model.vocab, word2vec_model.vectors):
        word2vec_dict[word] = vector
    word_embedding_2dlist = [[]] * vocab_size  # create an empty word_embedding list.
    word_embedding_2dlist[0] = np.zeros(FLAGS.embed_size)  # assign empty for first word:'PAD'
    bound = np.sqrt(6.0) / np.sqrt(vocab_size)  # bound for random variables.
    count_exist = 0;
    count_not_exist = 0
    for i in range(1, vocab_size):  # loop each word
        word = vocabulary_index2word[i]  # get a word
        embedding = None
        try:
            embedding = word2vec_dict[word]  # try to get vector:it is an array.
        except Exception:
            embedding = None
        if embedding is not None:  # the 'word' exist a embedding
            word_embedding_2dlist[i] = embedding;
            count_exist = count_exist + 1  # assign array to this word.
        else:  # no embedding for this word
            word_embedding_2dlist[i] = np.random.uniform(-bound, bound, FLAGS.embed_size);
            count_not_exist = count_not_exist + 1  # init a random value for the word.
    word_embedding_final = np.array(word_embedding_2dlist)  # covert to 2d array.
    word_embedding = tf.constant(word_embedding_final, dtype=tf.float32)  # convert to tensor
    t_assign_embedding = tf.assign(textCNN.Embedding,word_embedding)  # assign this value to our embedding variables of our model.
    sess.run(t_assign_embedding);
    print("word. exists embedding:", count_exist, " ;word not exist embedding:", count_not_exist)
    print("using pre-trained word emebedding.ended...")

# 在验证集上做验证，报告损失、精确度