word2vec.py 958 Bytes
import re
# from gensim.models import word2vec

def simple_word2vec(text):
    clean_text = text.strip()
    text_len = len(clean_text)

    digit_num = 0
    en_num = 0
    cn_num = 0
    space_num = 0
    other_num = 0
    for char in clean_text:
        if char.isdigit():
            digit_num += 1
        elif re.match(r'[A-Za-z]', char):
            en_num += 1
        elif char.isspace():
            space_num += 1
        elif re.match(r'[\u4e00-\u9fa5]', char):
            cn_num += 1
        else:
            other_num += 1
    
    vec = [text_len/100,
           cn_num/text_len,
           en_num/text_len,
           digit_num/text_len,
           # space_num/text_len,
           other_num/text_len,
           ]

    # print(text)
    # print(clean_text)
    # print('-------------')
    # print(en_num)
    # print(cn_num)
    # print(digit_num)
    # print(space_num)
    # print(other_num)
    # print('-------------')

    return vec