twi_wordcloud.py

import sys, getopt
import csv, re, MeCab
from os import path
from wordcloud import WordCloud
import matplotlib.pyplot as plt

def clean_tweet(tweet):
    reply = r'@[A-Za-z0-9\_\-]+'
    hashtag = r'#\w+'
    link = r'\w+:\/\/\S+'
    rt = r'RT'
    punc = r'[:\[\]ー,\.、。\']'
    face = r'm?\\?＼?₍{0,2} ?\([^a-np-zA-Zぁ-ふほ-んァ-ン一-龥]{2,14}\) ?\/?／?₎{0,2}m?.?☞?|（[^a-np-zA-Zぁ-んァ-ン一-龥]{2,12}）|\(.3\[.{1,5}.|・ω.{1,2}|¯\\_\(ツ\)_/¯|⊂.*[⊃₎]|_\(:3.*\)_|.*\)\"\"|.*ฅ|ʕ.*ʔ| ҉.*\*҉|\(.*੭ु| *＿人*＿| *￣Y.*￣|【| 】|＞.*＜|>.*<'
    regex = reply+'|'+hashtag+'|'+link+'|'+rt+'|'+punc+'|'+face
    tweet = re.sub(regex, '', tweet)
    return tweet

def mecab_analysis(tweet):
    tagger = MeCab.Tagger('-Ochasen -d /usr/local/lib/mecab/dic/ipadic')
    for chunk in tagger.parse(clean_tweet(tweet)).splitlines()[:-1]:
        surface,_,_,feature,_,_ = chunk.split('\t')
        if bool(re.search(r'^[名形動]',feature)):
            return surface

def parse_csv(csvfile):
    string = ''
    with open (csvfile, newline='') as f:
        reader = csv.reader(f)
        for row in reader:
            string = string + str(mecab_analysis(row[5])) + ' '
    return string

def draw_cloud(txt):
    font = '/Library/Fonts/ヒラギノ丸ゴ ProN W4.ttc'
    wordcloud = WordCloud(
        font_path=font,
        width=900,
        height=500
        ).generate(txt)
    plt.figure()
    plt.imshow(wordcloud)
    plt.axis("off")
    plt.show()

def main(argv):
    inputfile = ''

    try:
        opts, args = getopt.getopt(argv, "hi:o:", ["ifile="])
    except getopt.GetoptError:
        print ('python3 twitter_analysis.py -i <inputfile>')
        sys.exit(2)

    for opt, arg in opts:
        if opt in ("-i"):
            inputfile = arg

    draw_cloud(parse_csv(inputfile))

if __name__ == "__main__":
    print("starting...")
    main(sys.argv[1:])

$ brew install mecab
$ brew install mecab-python3
$ brew install mecab-ipdadic
$ git clone --depth 1 https://github.com/neologd/mecab-ipadic-neologd.git
$ pip3 install wordcloud

実行

$ python3  twi_wordcloud.py -i tweets.csv