Gordo

- Support building models without scoring/cross val (326)
- Fix issue where the serializer drops parameters to keras (333)
- Refactor ML Server into modular model views (288)
- Rename auto encoders .transform() -> .predict() (288)
- Upgrade scikit-learn ~=0.21

Note: This depends on a compatible version of gordo-infrastructure, the soon-to-be-released 0.24.0

- Replace generate_window with keras functionality, speeding up `predict` and improving `fit` (299)
- Ensure model-config in builder is fully expanded (313)
- Fix setup.py license and supported python versions (314)

Breaking change:
- Add `name` as a mandatory input to model building, and insert the name into metadata (291)

Other:
- Upgrade several test-dependencies (309)
- Add major and minor versions to model key calculation, so now models are rebuilt with new versions of gordo-components (312)

Minor version with a few fixes:
- Client should also retry on BadRequest errors (295)
- Upgrade several dependencies in the docker builds (307)
- We now push pypi packages on releases.

This is a bugfix release on 0.20.1 with the following changes:
- Uses the proper TimeoutError in the client (292)
- Improved logging in the server (289)

Breaking change:
- Rename Watchman's "metadata" top-level key to "endpoint-metadata" (249)

Other:
- Watchman: set endpoint healthy via metadata response instead of seperate request (279)
- Add resampling_endpoint up to which data will be forward filled (269)
- Client: Support retrying failed batches for POST requests
- Client: Improve logging (280)