Separate the query from the request URL, and use only the path as the S3 key

Currently, a request that includes a query string will fail, as the entire URL, including the query, will be used as the S3 key.

Instead, if there is a query, it should be excluded from the key:

url: /index.html?foo=bar

should use index.html as the S3 key. The query object should be preserved to be used at some later date.