I'm brand new to Scrapy. I have learned how to use response.css() for reading specific aspects from a web page, and am avoiding learning the xpath system. It seems to do the exact same thing, but in a different format (correct me if I'm wrong)

The site I'm scraping has long paragraphs of text, with an occasional linked text right in the middle. This sentence with a link to a picture of a dog is an example. I'm not sure if there is a way to have a spider read the text, with links in place (I've only been using response.css("p::text").extract())

Is there a way, using CSS (preferably) or xpath that I can grab all text in the paragraphs including the link-embedded text, without moving the links or link-text out of the sentence? The wording is difficult on this so apologies if I need to re-explain or give an example.

edit: some clarification is needed, this was poorly explained initially. A statement in this webpage can look like: <p>My sentence has a <a href="https://www.google.com">link to google</a> in it.</p> But when you use response.css("p::text").extract(), that sentence would show up as the list ["My sentence has a ","in it."], completely negating the text in the link. My goal is to get: ["My sentence has a link to google in it."]

2 Answers

vezunchik On Best Solutions

You can try to extract text with this expression:

>>> txt = """<p>My sentence has a <a href="https://www.google.com">link to google</a> in it.</p>"""
>>> from scrapy import Selector
>>> sel = Selector(text=txt)
>>> sel.css('p ::text').extract()
[u'My sentence has a ', u'link to google', u' in it.']
>>> ' '.join(sel.css('p ::text').extract())
u'My sentence has a  link to google  in it.'

Or, for example, use w3lib.html library to clean html tags from your response. In this way:

from w3lib.html import remove_tags
with_tags = response.css("p").get()
clean_text = remove_tags(with_tags)

But first variant looks shorter and more readable.

Gallaecio On

Use html-text after extracting the whole paragraph:

from html_text import extract_text

for paragraph in response.css('p'):
    html = paragraph.get()
    text = extract_text(html)