使用 ne_chunks 提取全名

Question

这里是新手。我正在尝试使用以下代码提取人员和组织的全名。

def get_continuous_chunks(text):
    chunked = ne_chunk(pos_tag(word_tokenize(text)))
    continuous_chunk = []
    current_chunk = []
    for i in chunked:
        if type(i) == Tree:
            current_chunk.append(' '.join([token for token, pos in i.leaves()]))
            if current_chunk:
                named_entity = ' '.join(current_chunk)
                if named_entity not in continuous_chunk:
                    continuous_chunk.append(named_entity)
                    current_chunk = []
                else:
                    continue
                return continuous_chunk

            
>>> my_sent = "Toni Morrison was the first black female editor in fiction at Random House in New York City."
>>> get_continuous_chunks(my_sent)
['Toni']

如您所见，它只返回第一个专有名词。不是全名，也不是字符串中的任何其他专有名词。

我做错了什么？

Answer 1

这是一些工作代码。

最好的办法是单步执行代码并在不同的地方放置大量打印语句。您将看到我在哪里打印了您正在迭代的项目的 type() 和 str() 值。我发现这有助于我想象并更多地思考我正在编写的循环和条件，如果我能看到它们被列出的话。

此外，哎呀，我无意中将所有变量命名为“连续”而不是“连续”……不确定为什么……连续可能更准确

代码：

from nltk import ne_chunk, pos_tag, word_tokenize
from nltk.tree import Tree


def get_continuous_chunks(text):
    chunked = ne_chunk(pos_tag(word_tokenize(text)))
    current_chunk = []
    contiguous_chunk = []
    contiguous_chunks = []

    for i in chunked:
        print(f"{type(i)}: {i}")
        if type(i) == Tree:
            current_chunk = ' '.join([token for token, pos in i.leaves()])
            # Apparently, Tony and Morrison are two separate items,
            # but "Random House" and "New York City" are single items.
            contiguous_chunk.append(current_chunk)
        else:
            # discontiguous, append to known contiguous chunks.
            if len(contiguous_chunk) > 0:
                contiguous_chunks.append(' '.join(contiguous_chunk))
                contiguous_chunk = []
                current_chunk = []

    return contiguous_chunks

my_sent = "Toni Morrison was the first black female editor in fiction at Random House in New York City."


print()
contig_chunks = get_continuous_chunks(my_sent)
print(f"INPUT: My sentence: '{my_sent}'")
print(f"ANSWER: My contiguous chunks: {contig_chunks}")

执行：

(venv) [ttucker@zim Whosebug]$ python contig.py 

<class 'nltk.tree.Tree'>: (PERSON Toni/NNP)
<class 'nltk.tree.Tree'>: (PERSON Morrison/NNP)
<class 'tuple'>: ('was', 'VBD')
<class 'tuple'>: ('the', 'DT')
<class 'tuple'>: ('first', 'JJ')
<class 'tuple'>: ('black', 'JJ')
<class 'tuple'>: ('female', 'NN')
<class 'tuple'>: ('editor', 'NN')
<class 'tuple'>: ('in', 'IN')
<class 'tuple'>: ('fiction', 'NN')
<class 'tuple'>: ('at', 'IN')
<class 'nltk.tree.Tree'>: (ORGANIZATION Random/NNP House/NNP)
<class 'tuple'>: ('in', 'IN')
<class 'nltk.tree.Tree'>: (GPE New/NNP York/NNP City/NNP)
<class 'tuple'>: ('.', '.')
INPUT: My sentence: 'Toni Morrison was the first black female editor in fiction at Random House in New York City.'
ANSWER: My contiguous chunks: ['Toni Morrison', 'Random House', 'New York City']

我也不太清楚你到底在找什么，但从描述来看，好像是这样。

使用 ne_chunks 提取全名

Extracting full names with ne_chunks

python-3.x

chunks

named-entity-recognition

代码：

执行：