Spacy：计算每个句子中特定标记的出现次数

Question

我想使用 spacy 计算语料库中每个句子的标记 和 的出现，并将每个句子的结果附加到列表中。到现在为止，下面的代码返回关于和的总数（整个语料库）。

3 个句子的示例/期望输出：['1', '0', '2'] 当前输出：[3]

doc = nlp(corpus)
nb_and = []
for sent in doc.sents:
    i = 0
    for token in sent:
        if token.text == "and":
            i += 1
            nb_and.append(i)

Answer 1

每句处理完后需要在nb_and后面追加i：

for sent in doc.sents:
    i = 0
    for token in sent:
        if token.text == "and":
            i += 1
    nb_and.append(i)

测试代码：

import spacy
nlp = spacy.load("en_core_web_trf")
corpus = "I see a cat and a dog. None seems to be unhappy. My mother and I wanted to buy a parrot and a tortoise."
doc = nlp(corpus)
nb_and = []
for sent in doc.sents:
    i = 0
    for token in sent:
        if token.text == "and":
            i += 1
    nb_and.append(i)

nb_and
# => [1, 0, 2]