sudachipyを試してみた - Qiita

tokenizer_obj = dictionary.Dictionary(dict="full").create() mode = tokenizer.Tokenizer.SplitMode.C doc = [] for row in range(len(df)): t = tokenizer_obj.tokenize(df["text"][row], mode) d = [m.normalized_form() for m in t if m.part_of_speech()[0] in ["名詞", "動詞"]] doc.append(d) docs = pd.array([" ".join(doc[i]) for i in range(len(doc))]) print(docs) #<StringArray> #['メロス 激怒 為る', #'邪知 暴虐 王 除く 成る 決意 為