Google Bucket File System Support

[luke14free]

Along with the other filesystem requests, it would be amazing to be able to use papermill with Google Buckets filesystems. I'll try to contribute to this myself as well.

[MSeal]

Awesome! Do feel free to post the WIP PR if you get stuck or want input part way through.

[bnaul]

I'm interested in implementing this; the same thing for pandas was straightforward pandas-dev/pandas#22704. Does it make sense to do so before #257? I'm personally not thrilled about installing all the Azure and S3 IO tools to use papermill and I can imagine other people feeling the same way about gcsfs.

[MSeal]

If you get a PR open for this feature I can get #257 implemented so we can start isolating dependencies and we can hold off merging until it's in place. The dependency isolation is needed for papermill 1.0 so I need to do it regardless.

[gogasca]

We are planning to launch Blog including a workaround to run Papermill using a GCP Bucket. Please take a look at this: https://blog.kovalevskyi.com/how-to-submit-jupyter-notebook-for-overnight-training-on-gcp-4ce1b0cd4d0d ideally we would like to have native support. Not sure what is the status so far. Interested in contributing too.