Add history query API and persistent implementation

[amitschang]

The history query API supports some basic query parameters and specifies the output shape that should be returned using models so clients can get schema. This sets the interface in library and the web api call.

The default implementation has been built-out to be persistent based on json files, and with support for "experiment" isolation. It uses hive-style partitioning to enable efficient query for long-running experiments and simpler management of data (but does not include any kind of cleanup itself).

---

Notes:

• The history query interface is fairly concrete, but I think it makes sense for the kind of data. One future option to support a more arbitrary pluggable query is to add another endpoint and an optional query model to history servers - but this might not have strong case.
• All history can be paged through via the query parameters, but it may be worthwhile to have an option to dump all history via a stream or chunked response - but that may also be unnecessary as files can be gotten directly from pi, or an appropriate tool could do the paging...

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 95.55556% with 6 lines in your changes missing coverage. Please review.	Files	Patch %	Lines
evolver/history/standard.py	92.20%	6 Missing :warning:

:loudspeaker: Thoughts on this report? Let us know!