Flickr-url-parser

Add support for recognising URLs as the Flickr homepage.

There are lots of varieties of homepage URL that appear in e.g. links from Wikimedia Commons, and now they can be recognised:

pycon
>>> parse_flickr_url("www.flickr.com")
{"type": "homepage"}

* Fix a bug where the URL parser could throw an IndexError on URL fragments or empty strings.

* Add a new function `find_flickr_urls_in_text` which can be used to find Flickr URLs in a block of arbitrary text.
Example:

pycon
>>> text = """
... This is the help page: https://www.flickr.com/help
...
... This is a user: https://www.flickr.com/photos/mariakallin/
... """
>>> find_flickr_urls_in_text(text)
['https://www.flickr.com/help', 'https://www.flickr.com/photos/mariakallin/']


This is useful for text analysis.

Add support for parsing more varieties of URL, based on those seen in the Wikimedia Commons snapshots, including:

* Old-style photo URLs that use `/photo/` instead of `/photos/`, e.g. `http://flickr.com/photo/17277074N00/2619974961`
* Photo URLs that use the `/photo_zoom.gne` path and similar `.gne` paths, e.g. `https://www.flickr.com/photo_zoom.gne?id=196155401&size=m`
* A wide variety of `static.flickr.com`-like URLs; URLs that point to raw JPEGs rather than the photo description page

* Expand the support for parsing static URLs, e.g. `http://farm1.static.flickr.com/82/241abc183_dd0847d5c7_o.jpg`

* Expand the support for parsing static URLs, e.g. `https://photos12.flickr.com/16159487_3a6615a565_b.jpg`