hatebu_comment_mecab.py

import urllib.parse
import urllib.request
import json
import pdb # pdb.set_trace()

# pip install Janome
# from janome.tokenfilter import POSKeepFilter
# from janome.tokenizer import Tokenizer
# from janome.analyzer import Analyzer

# pip install mecab-python3
import MeCab

mecab = MeCab.Tagger ('-d /usr/local/lib/mecab/dic/mecab-ipadic-neologd')

base_url    = 'http://b.hatena.ne.jp/entry/jsonlite/?url=https%3A%2F%2F'
article_url = "anond.hatelabo.jp/20180604174309"

def request_hatena():
    res      = urllib.request.urlopen(base_url + article_url)
    body     = res.read().decode('utf-8')
    print(body)

def debug_print():
    f = open('a.json','r')
    json_data = json.load(f)

    for row in json_data['bookmarks']:
        print(row['comment'])

    f.close()

def analyzer_json():
    f = open('a.json','r')
    json_data = json.load(f)
    manga = {}

    for row in json_data['bookmarks']:
        node = mecab.parseToNode(row['comment'])
        while node:
            word = node.surface
            pos = node.feature.split(",")[1]
            if word is None or pos != '固有名詞':
                node = node.next
                continue

            # print('{0} , {1}'.format(word, pos))
            try:
                manga[word] += 1
            except KeyError as e:
                manga[word] = 1

            node = node.next

    for k, v in sorted(manga.items(), key=lambda x:x[1], reverse=True):
       print(k, v)
    f.close()

analyzer_json()