Add anchor links to versions in the changelog

[tswast]

Uses Sphinx syntax for cross-reference labels.

[max-sixty]

Thanks @tswast

[tswast]

Of course!

For posterity, I added these anchors so that I could link to them from the Pandas docs at pandas-dev/pandas#21628

[max-sixty]

Super!

I've been out of action on OSS recently given work, and heading on vacation for a couple of weeks. Hopefully I'll have more time in the Summer. Let me know if you see pandas-gbq issues that you think would be worth focusing on. Would be good for me to get something in re bulk download.

[tswast]

No worries. These are the top priorities on my mind for pandas-gbq:

1. Auth. pandas-gbq auth proposal #161 Basically, add a client parameter plus a couple public helper methods for auth and deprecate all the other auth-related parameters.
2. to_gbq configuration argument. Take in a job configuration dictionary, just like read_gbq does. It might make sense to officially deprecate the chunksize argument at that time, too. BUG: Add support to replace partitions in date-partitioned tables (#43) #124 (comment)
3. Delegate more to google-cloud-bigquery document pandas-gbq vision and roadmap #149 (comment)

Bulk download will be an interesting one to investigate. It does depend on GCS, so it'll likely have to be a separate read_gbq_bulk method, since we need a scratch bucket to work with. (Plus we'll have to update the scopes we use for authentication.)