wordCloud

从《荷塘月色》提词绘制词云图

mask_img

data

 code

import jieba
import numpy as np
from PIL import Image
from wordcloud import WordCloud


def gen_wordcloud(text_file, mask_img):
    """
    Parameters
    ----------
    text_file :一个文本文件名，函数对此文件生成词云
    mask_img : 一个.jpg文件名，词云mask的图像文件
    Returns
    -------
    None.
    函数读取text_file,以mask_img为背景生成词云
    """
    mask = np.array(Image.open(mask_img))
    # 从此处继续书写函数gen_wordcloud的代码
    with open(text_file, encoding='utf-8') as f:
        text = f.read()

    raw = list(jieba.cut(text))

    word = {}
    for i in raw:
        if len(i) > 1:
            if i not in word:
                word[i] = 1
            else:
                word[i] += 1

    wc = WordCloud(font_path='msyh.ttc', background_color='white', max_words=120, mask=mask, width=1000, height=800)
    wc.generate_from_frequencies(word)
    wc.to_file('词云.png')


if __name__ == '__main__':
    gen_wordcloud('./荷塘月色.txt', './cloud.jpg')