from knock30 import neko_mecab list_mecab = [] for line in neko_mecab(): list_mecab.append(line) for i, line in enumerate(list_mecab): if line['surface'] == 'の': if list_mecab[i - 1]['pos'] == '名詞' and list_mecab[i + 1]['pos'] == '名詞': print(list_mecab[i - 1]['surface'] + line['surface'] + list_mecab[i + 1]['surface'])
from collections import defaultdict from knock30 import neko_mecab import matplotlib.pyplot as plt from matplotlib.font_manager import FontProperties words_mecab = defaultdict(lambda:0) count_of_words = 0 l = [] n = [] for line in neko_mecab(): for word in line['surface']: if word not in words_mecab: count_of_words += 1 words_mecab[word] += 1 for key, value in reversed(sorted(words_mecab.items(), key=lambda x:x[1])): l.append(value) n.append(key) print(l) plt.hist(l,50, range=(1,50)) plt.show()