CODEWITHSUNDEEP
pythonhtmlxpathcss-selectorsscrapy
Dimitry
Dimitry

Reputation: 369

Scrapy. Extract html from div without wrapping parent tag

I use scrapy to crawl a website.

I want to extract contents of certain div.

<div class="short-description">
{some mess with text, <br>, other html tags, etc}
</div>

loader.add_xpath('short_description', "//div[@class='short-description']/div")

By that code I get what I need but result includes wrapping html (<div class="short-description">...</div>)

How to get rid of that parent html tag?

Note. Selector like text(), node() cannot help me, because my div contains <br>, <p>, other divs, etc., whitespaces, and I need to keep them.

Upvotes: 3

Views: 3983

Answers (2)

Steven Almeroth
Steven Almeroth

Reputation: 8202

Try node() in combination with Join():

loader.get_xpath('//div[@class="short-description"]/node()', Join())

and the results look something like:

>>> from scrapy.contrib.loader import XPathItemLoader
>>> from scrapy.contrib.loader.processor import Join
>>> from scrapy.http import HtmlResponse
>>>
>>> body = """
...     <html>
...         <div class="short-description">
...             {some mess with text, <br>, other html tags, etc}
...             <div>
...                 <p>{some mess with text, <br>, other html tags, etc}</p>
...             </div>
...             <p>{some mess with text, <br>, other html tags, etc}</p>
...         </div>
...     </html>
... """
>>> response = HtmlResponse(url='http://example.com/', body=body)
>>>
>>> loader = XPathItemLoader(response=response)
>>>
>>> print loader.get_xpath('//div[@class="short-description"]/node()', Join())

            {some mess with text,  <br> , other html tags, etc}
             <div>
                <p>{some mess with text, <br>, other html tags, etc}</p>
            </div>
             <p>{some mess with text, <br>, other html tags, etc}</p>
>>>
>>> loader.get_xpath('//div[@class="short-description"]/node()', Join())
u'\n            {some mess with text,  <br> , other html tags, etc}\n
   <div>\n         <p>{some mess with text, <br>, other html tags, etc}</p>\n
   </div> \n     <p>{some mess with text, <br>, other html tags, etc}</p> \n'

Upvotes: 2

perreal
perreal

Reputation: 98108

hxs = HtmlXPathSelector(response)
for text in hxs.select("//div[@class='short-description']/text()").extract(): 
    print text

Upvotes: 1

Related Questions