Metadata-parser

- cleaned up some code

- realized that some servers return gzip content, despite not advertising that this client accepts that content ; fixed by using some ideas from mark pilgrim's feedparser. metadata_parser now advertises gzip and zlib, and processes it as needed

- fixed a bug that prevented toplevel directories from being parsed

- made redirect/masked/shortened links have better dereferenced url support

- Wrapped title tag traversal with an AttributeException try block
- Wrapped canonical tag lookup with a KeyError try block, defaulting to 'href' then 'content'
- Added support for `url_actual` and `url_info` , which persist the data from the urllib2.urlopen object's `geturl()` and `info()`
- `get_discrete_url` and `absolute_url` use the underlying url_actual data
- added support for passing data and headers into urllib2 requests

Initial Release