Simple improvements to URL normalization

[sylvinus]

There are many simple things we could do to improve our normalized URLs and avoid duplicates:

• [ ] Remove known session or marketing trackers (PHPSESSID, utm_*, ..)
• [x] Remove default port numbers (80 for http, 443 for https)
• [ ] Remove wwwNN subdomains (we already remove www)
• [ ] Sort query string params ?

Some good ideas there: https://github.com/iipc/webarchive-commons/tree/master/src/main/java/org/archive/url https://github.com/rajbot/surt/tree/master/surt

Code for this should be done in https://github.com/commonsearch/cosr-back/blob/master/cosrlib/url.py. Exhaustive unit tests would be great!