Scraping

Scraping (also Screen Scraping) is the process of extracting data out of something.

In the course, we have used the library BeautifulSoup to manipulate HTML pages in Python.

Other interesting libraries to consider:

• Mechanize in essence simulates a browser in Python, that can "remember" things (like cookies / sessions) between pages
• lxml which can apparently deal with "mal-formed" HTML and quickly convert them to xml trees
• html5lib

Resources:

• http://us.pycon.org/2009/tutorials/schedule/2AM8/
• http://scrapy.org/ Python framework for custom scrapers

See Extracting parts of an HTML document and other recipes in the Category:Cookbook

Scraping HTML in Python with html5lib + css selectors (2022)

html5lib is the modern python library for parsing "HTML in the wild" ... it's basically guaranteed to handle any HTML content and will correct broken tags, etc. to make sure whatever you load is "well-formed". Alternatively, you can use a library like BeautifulSoup.

   pip install html5lib elementpath cssselect

from xml.etree import ElementTree as ET
import html5lib

def textContent (elt):
    if elt.text != None:
        ret = elt.text
    else:
        ret = u""
    return ret + u"".join([ET.tostring(x, method="text", encoding="utf8").decode("utf-8") for x in elt])

with open("index.html") as fin:
    t = html5lib.parse(fin, namespaceHTMLElements=False)

print (textContent(t.find(".//p")))

from cssselect import HTMLTranslator
import elementpath

def querySelector (t, selector):
    for elt in elementpath.select(t, HTMLTranslator().css_to_xpath(selector)):
        return elt

def querySelectorAll (t, selector):
    for elt in elementpath.select(t, HTMLTranslator().css_to_xpath(selector)):
        yield elt

for p in querySelectorAll("p.test"):
    print (textContent(p))