Is it possible for Scrapy to get plain text from raw HTML data?

Question

For example:

scrapy shell http://scrapy.org/
content = hxs.select('//*[@id="content"]').extract()[0]
print content

Then, I get the following raw HTML code:

<div id="content">


  <h2>Welcome to Scrapy</h2>

  <h3>What is Scrapy?</h3>

  <p>Scrapy is a fast high-level screen scraping and web crawling
    framework, used to crawl websites and extract structured data from their
    pages. It can be used for a wide range of purposes, from data mining to
    monitoring and automated testing.</p>

  <h3>Features</h3>

  <dl>

    <dt>Simple</dt>
    <dt>
    </dt>
    <dd>Scrapy was designed with simplicity in mind, by providing the features
      you need without getting in your way
    </dd>

    <dt>Productive</dt>
    <dd>Just write the rules to extract the data from web pages and let Scrapy
      crawl the entire web site for you
    </dd>

    <dt>Fast</dt>
    <dd>Scrapy is used in production crawlers to completely scrape more than
      500 retailer sites daily, all in one server
    </dd>

    <dt>Extensible</dt>
    <dd>Scrapy was designed with extensibility in mind and so it provides
      several mechanisms to plug new code without having to touch the framework
      core

    </dd>
    <dt>Portable, open-source, 100% Python</dt>
    <dd>Scrapy is completely written in Python and runs on Linux, Windows, Mac and BSD</dd>

    <dt>Batteries included</dt>
    <dd>Scrapy comes with lots of functionality built in. Check <a
        href="http://doc.scrapy.org/en/latest/intro/overview.html#what-else">this
      section</a> of the documentation for a list of them.
    </dd>

    <dt>Well-documented &amp; well-tested</dt>
    <dd>Scrapy is <a href="/doc/">extensively documented</a> and has an comprehensive test suite
      with <a href="http://static.scrapy.org/coverage-report/">very good code
        coverage</a></dd>

    <dt><a href="/community">Healthy community</a></dt>
    <dd>
      1,500 watchers, 350 forks on Github (<a href="https://github.com/scrapy/scrapy">link</a>)<br>
      700 followers on Twitter (<a href="http://twitter.com/ScrapyProject">link</a>)<br>
      850 questions on StackOverflow (<a href="http://stackoverflow.com/tags/scrapy/info">link</a>)<br>
      200 messages per month on mailing list (<a
        href="https://groups.google.com/forum/?fromgroups#!aboutgroup/scrapy-users">link</a>)<br>
      40-50 users always connected to IRC channel (<a href="http://webchat.freenode.net/?channels=scrapy">link</a>)
    </dd>

    <dt><a href="/support">Commercial support</a></dt>
    <dd>A few companies provide Scrapy consulting and support</dd>

    <p>Still not sure if Scrapy is what you're looking for?. Check out <a
        href="http://doc.scrapy.org/en/latest/intro/overview.html">Scrapy at a
      glance</a>.

    </p>
    <h3>Companies using Scrapy</h3>

    <p>Scrapy is being used in large production environments, to crawl
      thousands of sites daily. Here is a list of <a href="/companies/">Companies
        using Scrapy</a>.</p>

    <h3>Where to start?</h3>

    <p>Start by reading <a href="http://doc.scrapy.org/en/latest/intro/overview.html">Scrapy at a glance</a>,
      then <a href="/download/">download Scrapy</a> and follow the <a
          href="http://doc.scrapy.org/en/latest/intro/tutorial.html">Tutorial</a>.


    </p></dl>
</div>

But I want to get plain text directly from scrapy.

I do not want to use any xPath selectors to extract the p, h2, h3... tags, since I am crawling a website whose main content is embedded into a table, tbody; recursively. It can be a tedious task to find the xPath.

Can this be implemented by a built in function in Scrapy? Or do I need external tools to convert it? I have read through all of Scrapy's docs, but have gained nothing.

This is a sample site which can convert raw HTML into plain text: http://beaker.mailchimp.com/html-to-text

Answer 1

Scrapy doesn't have such functionality built-in. html2text is what you are looking for.

Here's a sample spider that scrapes wikipedia's python page, gets first paragraph using xpath and converts html into plain text using html2text:

from scrapy.selector import HtmlXPathSelector
from scrapy.spider import BaseSpider
import html2text


class WikiSpider(BaseSpider):
    name = "wiki_spider"
    allowed_domains = ["www.wikipedia.org"]
    start_urls = ["http://en.wikipedia.org/wiki/Python_(programming_language)"]

    def parse(self, response):
        hxs = HtmlXPathSelector(response)
        sample = hxs.select("//div[@id='mw-content-text']/p[1]").extract()[0]

        converter = html2text.HTML2Text()
        converter.ignore_links = True
        print(converter.handle(sample)) #Python 3 print syntax

prints:

**Python** is a widely used general-purpose, high-level programming language.[11][12][13] Its design philosophy emphasizes code readability, and its syntax allows programmers to express concepts in fewer lines of code than would be possible in languages such as C.[14][15] The language provides constructs intended to enable clear programs on both a small and large scale.[16]

Answer 2

Another solution using lxml.html's tostring() with parameter method="text". lxml is used in Scrapy internally. (parameter encoding=unicode is usually what you want.)

See http://lxml.de/api/lxml.html-module.html for details.

from scrapy.spider import BaseSpider
import lxml.etree
import lxml.html

class WikiSpider(BaseSpider):
    name = "wiki_spider"
    allowed_domains = ["www.wikipedia.org"]
    start_urls = ["http://en.wikipedia.org/wiki/Python_(programming_language)"]

    def parse(self, response):
        root = lxml.html.fromstring(response.body)

        # optionally remove tags that are not usually rendered in browsers
        # javascript, HTML/HEAD, comments, add the tag names you dont want at the end
        lxml.etree.strip_elements(root, lxml.etree.Comment, "script", "head")

        # complete text
        print lxml.html.tostring(root, method="text", encoding=unicode)

        # or same as in alecxe's example spider,
        # pinpoint a part of the document using XPath
        #for p in root.xpath("//div[@id='mw-content-text']/p[1]"):
        #   print lxml.html.tostring(p, method="text")