Add statistics from response headers for query_multiple

[vossen-adobe]

The information is already available in the response headers. My goal with this PR is to add them to the QueryMultiple class as instance variables. This way, when a users or groups query is run (for example, in umapi.py in UST), the resulting object includes the meta info about results.

Here, I added the instance variables to track the following for each request made:

_total_count = x-total-count
_page_size = x-page-count
_page_count = x-current-page
_page_number = x-page-size

Since the info is already on the response, no extra work is needed but to track and set them. Using the resulting fields, it is possible to track the progress of a prolonged umapi run - which is especially useful in user sync tool, where runs often exceed several hours. In a separate PR, I will utilize this feature to add the progress logging in UST.

No new tests are added, but existing tests using mock response have been updated to test these fields as well. The travis failure is because of a dependency issue between pyyaml and python 3.4 - not related to these changes.

[adorton-adobe]

We should add a method to QueryMultiple that returns these new stats so we don't have to access private members of the object. Maybe stats() or something like that.

[vossen-adobe]

@adorton-adobe This is up to date and resolved