Add self.retry_after to SparkRateLimitError class

[sardoodledom]

If one sets the wait_on_rate_limit to True in their CiscoSparkAPI
class instantiation, they will hit an error telling them that
retry_after is not an attribute of the SparkRateLimitError class.

This is because the RestSession class checks for its existence
when it catches the error at line 282 in restsession.py

Error:
File "/home/daspano/../ciscosparkapi/restsession.py", line 282, in request
if self.wait_on_rate_limit and e.retry_after:
AttributeError: 'SparkRateLimitError' object has no attribute 'retry_after'

This commit adds that the retry_after attribute as an int, as my testing
shows the Retry-After result always being an integer in seconds.

Ex:
-----------------------------------Response------------------------------------
429 Too Many Requests
Date: Wed, 15 Nov 2017 21:47:00 GMT
Content-Type: application/json
Content-Length: 182
Connection: keep-alive
TrackingID: ROUTER_5A0CB5D2-B92F-01BB-7C3F-AC12D9227C3F
Cache-Control: no-cache
Retry-After: 194
Server: Redacted
l5d-success-class: 1.0
Via: 1.1 linkerd
content-encoding: gzip
Strict-Transport-Security: max-age=63072000; includeSubDomains; preload

[cmlccie]

@dlspano , thank you for catching this! This is a regression bug where I "cleaned up" and simplified the exception classes after the rate-limit handling was added. Have added a couple of consistency and syntax updates to your patch, and I'm running the test suit now. The changes should be merged in shortly.

[cmlccie]

@dlspano , the patch is included in v0.8.3, which is now available on PyPI.

Also, apparently the way I edited and merged in your changes caused your commit to get "merged" in with my updates to the patch, not showing you as the author of the original fix. I do apologize; this was not intentional. Please find another bug, update, enhancement or example that you would like to contribute to, and I'll make sure I don't squash your commit. I do appreciate your contributions to the package.

[cmlccie]

Scratch that... I didn't mess it up. Your contribution is there! 😎

[sardoodledom]

Thanks Chris! Our team is starting to do a lot of integration work with Spark's apis, creating bots, etc. If we run into anything that we can do to help out. We will.