Client-Side Web Programming

Libraries

utidylib and mxTidy -- Python interfaces to html tidy library to clean up HTML documents.
html5lib A HTML5-compliant library for parsing arbitarily-broken HTML to a range of tree formats including minidom, elementtree (including lxml) and BeautifulSoup
BeautifulSoup -- a permissive HTML parser.
Don't use HTMLParser on HTML that might be invalid! That way lies pain. Either clean it up (using tidy), or use a different parser.
urllib, urllib2, and httplib in the standard library.
ClientCookie, ClientForm, and Mechanize are higher-level libraries for writing a web client.
mechanoid a mechanize fork.
libxml2dom can parse HTML by employing libxml2's liberal HTML parser.