如何计算2个预定义单词之间的单词数量？

Question

<replace-add> that i dont know you know cause </replace-add> i could help you with <replace-del> that oh </replace-del> <replace-add> us </replace-add> thanks so i just set up a ride <replace-del> for </replace-del> <replace-add> from </replace-add> my daughter <replace-del> tenah dyer </replace-del> <replace-add> clear dire </replace-add>

How can i count the exact number of words between <replace-add> and </replace-add> in a text.

Answer 1

Without using any libraries:

def get_tag_indexes(text, tag, start_tag):
    tag_indexes = []
    start_index = -1

    while True:
        start_index = text.find(tag, start_index + 1)

        if start_index != -1:
            if start_tag:
                tag_indexes.append(start_index + len(tag))
            else:
                tag_indexes.append(start_index)
        else:
            return tag_indexes

text = """<replace-add>that i dont know you know cause</replace-add> i could help you with <replace-del>that oh</replace-del> <replace-add>us</replace-add> thanks so i just set up a ride <replace-del>for</replace-del> <replace-add>from</replace-add> my daughter <replace-del>tenah dyer</replace-del> <replace-add>clear dire</replace-add>"""

tag_starts = get_tag_indexes(text, "<replace-add>", True)
tag_ends = get_tag_indexes(text, "</replace-add>", False)

for start, end in zip(tag_starts, tag_ends):
    words = text[start:end].split()
    print "{} words - {}".format(len(words), words)

Giving you:

7 words - ['that', 'i', 'dont', 'know', 'you', 'know', 'cause']
1 words - ['us']
1 words - ['from']
2 words - ['clear', 'dire']

This uses a function to return a list of the locations of any given text. This can then be used to extract the text between two tags.

---

As an alternative approach, this could actually also be done using beautifulsoup:

from bs4 import BeautifulSoup

text = """<replace-add>that i dont know you know cause</replace-add> i could help you with <replace-del>that oh</replace-del> <replace-add>us</replace-add> thanks so i just set up a ride <replace-del>for</replace-del> <replace-add>from</replace-add> my daughter <replace-del>tenah dyer</replace-del> <replace-add>clear dire</replace-add>"""
soup = BeautifulSoup(text, "lxml")

for block in soup.find_all('replace-add'):
    words = block.text.split()
    print "{} words - {}".format(len(words), words)

Answer 2

Depending on how trusted the source is, you could do two things. Given that

source = """<replace-add>that i dont know you know cause</replace-add> i could help you with <replace-del>that oh</replace-del> <replace-add>us</replace-add> thanks so i just set up a ride <replace-del>for</replace-del> <replace-add>from</replace-add> my daughter <replace-del>tenah dyer</replace-del> <replace-add>clear dire</replace-add>"""

You could use regular expressions like so:

import re

from itertools import chain

word_pattern = re.compile(r"(?<=<replace-add>).*?(?=</replace-add>)")
re_words = list(chain.from_iterable(map(str.split, word_pattern.findall(source))))

This will only work if the source matches those tags exactly, with no attributes etc.

The other option in the standard library is HTML parsing:

from html.parser import HTMLParser

class MyParser(HTMLParser):
    def get_words(self, html):
        self.read_words = False
        self.words = []
        self.feed(html)
        return self.words

    def handle_starttag(self, tag, attrs):
        if tag == "replace-add":
            self.read_words = True

    def handle_data(self, data):
        if self.read_words:
            self.words.extend(data.split())

    def handle_endtag(self, tag):
        if tag == "replace-add":
            self.read_words = False


parser = MyParser()
html_words = parser.get_words(source)

This approach will be more reliable, and probably a little more efficient as it uses tools entirely focused on this task.

Now, doing

print(re_words)
print(html_words)

We get

['that', 'i', 'dont', 'know', 'you', 'know', 'cause', 'us', 'from', 'clear', 'dire']
['that', 'i', 'dont', 'know', 'you', 'know', 'cause', 'us', 'from', 'clear', 'dire']

(Of course, the len of this list is the number of words.)

If you strictly just require the number of words, you could just keep a running total and add the length of data.split to this total for each datum encountered.

If you really can't make any imports, you will either have to make some sacrifices, or have to implement your own regex engine/html parser. If this is a requirement of a homework assignment, really you should have shown some prior effort to posting the question.