Gee! These Aren't Roasted!
What regular expression would I use to extract <body> tags from a file?
These tags aren't always plain <body>; sometimes it's < body>, < body > (note the spaces), or < body someattributes >. What regexp(s) would I use in order to snag all opening body tags, regardeless of how they're formatted?
Thanks.
PS: I'm not sure if this matters, but I'll be using grep and sed.
Blue Ubuntu forums: http://userstyles.org/styles/2455 | Grey Ubuntu forums: http://userstyles.org/styles/3440 | Black Ubuntu forums: http://userstyles.org/styles/3992
Windows free since: 6/15/07
Banned
Parsing HTML by hand is IMHO sucker's game in general. If you have some simple static well-formed HTML, like in the case of <body> you may have luck, but in general - is better to use HTML parser, they solved most tricky corner cases for you.
I would use Python and ElementTree parser, but then I use Python everywhere I can get away with itYMMV
Last edited by pmasiar; February 4th, 2008 at 02:50 PM.
I Ubuntu, Therefore, I Am
how about providing a sample for people interested to play with and describing what you expect the final result to beOriginally Posted by Interestedinthepenguin
Iced Blended Vanilla Crème Ubuntu
Try...
Code:<\s*body[^>]*>
Banned
To get feeling what you might volunteered for if you want to parse HTML by hand, read about why dot-star is sometimes pronounced death-star: read Death to Dot Star! at PerlMonk
Iced Blended Vanilla Crème Ubuntu
I still stand by my regex! =O
Grande Half-n-Half Cinnamon Ubuntu
With that regex, <bodywhatever> is matched too (which is incorrect).
This fixes that specific flaw:
(note the ending i so that the regex is case insensitive)Code:/<\s*body(\s*>|\s+[^>]+>)/i
Handling of incorrectly formed HTML with a plain > inside an attribute value (it should be the > entity, but all browsers will happily parse it anyway*) is left as an exercise for the reader.
Example: <body attr="hello > world">
Handling of correctly formed XHTML that (for some obscure reason) uses a namespace prefix is left, too, as an exercise for the reader.
Example of such a case:
Code:<?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> <foo:html xmlns:foo="http://www.w3.org/1999/xhtml" xml:lang="en" foo:lang="en"> <foo:head> <foo:title>Tricky case</foo:title> </foo:head> <foo:body> Document contents... </foo:body> </foo:html>
These are definitely the kind of "tricky corner cases" that pmasiar was evoking, and the very reason why you should use a dedicated HTML parser (not an XML parser, mind you, as my first example would then not be parsed correctly despite the fact that all browsers will accept it*).
(*): if the browser thinks it's dealing with plain old HTML ; of course in XHTML mode it will fail to validate.
Still standing?
Last edited by aks44; February 4th, 2008 at 06:02 PM. Reason: typo + clarification
Not even tinfoil can save us now...
Ubuntu addict and loving it
RE is great for some situations, it just doesn't do well with hierarchical data like XML. You can parse simple things out of XML using RE, but XML is meant to be represented in a hierarchical / tree-like structure. I would use an XML parser as pmasiar suggested, however for true HTML I would probably use BeautifulSoup since it can handle badly formed HTML (like most of the internet).
Iced Blended Vanilla Crème Ubuntu
T_T I give up.
Grande Half-n-Half Cinnamon Ubuntu
While we're at it, I'll push the point a little farther...
It's quite easy to write a regex that checks if a string is a well-formed e-mail address, right?
Here's the beast, have fun!
http://code.iamcal.com/php/rfc822/full_regexp.txt
EDIT: this is a PHP / PCRE regex (Perl Compatible Regular Expression), other regex engines may require a different one.
Last edited by aks44; February 4th, 2008 at 03:50 PM.
Not even tinfoil can save us now...
Posting Permissions
Adv Reply
Bookmarks