repository_url key should normalize against minor differences

[DuaneOBrien]

Nearly identical repositories can result in data fragmentation. Sample:

  {
    "repository_url": "http://tomcat.apache.org",
    "score": 3793
  },
  {
    "repository_url": "https://tomcat.apache.org/",
    "score": 3293
  },
  {
    "repository_url": "http://tomcat.apache.org/",
    "score": 12
  }

It seems like we could strip the protocol and any trailing slash or whitespace characters and reduce this, while getting the same results.

[mjpitz]

The new API scheme introduced a concept of a ProviderURL that is intended to do this. Definitely seems like an easy thing we can fix right now.