::  Posts  ::  RSS  ::  ◂◂RSS  ::  Contact

Parsing HTML With Regular Expressions

February 22nd, 2013
tech, html  [html]
Perhaps you need to get some information out of HTML. Regular expressions look promising, but you get stuck so you ask for help. A typical response would be:
HTML parsing is not possible with regular expressions, since it depends on matching the opening and the closing tag. Regular expressions can only match regular languages but HTML is a context-free language.
This is true, but only in a sense so narrow it's useless. When someone asks about "regular expressions" they don't mean the restricted computer science kind but the implementations available in various programming languages. Those are much more powerful, capable of parsing HTML.

Here again we're using dangerously precise words. When someone says "HTML" they don't mean HTML as defined by the spec but HTML that they are likely to encounter in the real world. Which means you might need to handle monstrosities like "<i>italic <b>bold-italic</i> bold</b>" which have no parse tree but your browser still renders as italic bold-italic bold. [1]

But again with the precision! For typical tasks you don't need to 'parse' HTML, just extract some information from it. Regular expressions and other ad-hoc techniques can do that well and are likely to be less trouble in practice [2] than trying to use a parser.

(Because mod_pagespeed needs to make context-dependent changes to arbitrary web pages, regular expressions are not a good fit. You might think using the parsing code from a web browser would work well, but it turns out that browsers mix HTML-parsing with HTML-cleanup [1]. So pagespeed instead goes token-by-token, triggering callbacks for start and end tags.)

[1] Webkit turns <i>italic <b>bold-italic</i> bold</b> into <i>italic <b>bold-italic</b></i> <b>bold<b> at an early stage of interpreting the page.

[2] The advice to use an XML parser is just bad, as funny as the author is. XML parsers get to assume they're only going to be given valid XML and just reject anything that isn't. Extremely few pages are (or even try to be) XML, so your XML parser isn't going to be helpful.

Comment via: google plus, facebook, hacker news

Recent posts on blogs I like:

Absolute scale corrupts absolutely

The Internet has gotten too big. Growing up, I, like many computery people of my generation, was an idealist. I believed that better, faster communication would be an unmitigated improvement to society. "World peace through better communication,"…

via apenwarr August 19, 2019


At naptime Anna listens to recordings of novels recorded by Jeff’s grandmother. It is the main way she will know Winnie, as it is the main way I have ever known Winnie. Some of the recordings are missing parts, and Suzie often fills in the first few sente…

via The whole sky August 18, 2019

One Night Ultimate Werewolf

I really like One Night Ultimate Werewolf (ONUW), and recommend trying it out. The Daybreak expansion is also good (but I’d skip the later games in the series). App For those who’ve played a lot and wish the night took 20 seconds rather than a few minutes…

via The sideways view August 18, 2019

more     (via openring)

More Posts:

  ::  Posts  ::  RSS  ::  ◂◂RSS  ::  Contact