Gretel-client

**Full Changelog**: https://github.com/gretelai/gretel-python-client/compare/v0.11.0...v0.12.0

Updates
- Adds `gretel records get` command to download artifacts associated with a record handler
- `poll` helper is threadsafe
- Protocol can be omitted from endpoints during configuration (e.g. `api.gretel.cloud`)
- Projects can be deleted via the CLI: `gretel projects delete --name [NAME]`
- Improvements to report summaries

**Full Changelog**: https://github.com/gretelai/gretel-python-client/compare/0.10.0...v0.11.0

Updates

- First-class interfaces for using Gretel Evaluate! Which lets users create Synthetic Quality Score reports directly through the Gretel Service. See more details on our docs here: https://docs.gretel.ai/gretel.ai/evaluation-framework/overview

- Pandas DataFrame support for SDK interfaces. Now when creating models with the SDK, you can utilize Pandas DataFrames directly. In order to do this, Pandas must be installed separately from `gretel-client`. See example of this usage in our newly updated SDK docs: https://docs.gretel.ai/gretel.ai/cli-and-sdk-fundamentals/creating-models#creating-models-with-the-sdk

- When using the SDK, in addition to using the `poll()` function for tracking job status, you may now issue single `Model.refresh()` or `RecordHandler.refresh()` calls that will do a one time state update of a job's status from the Gretel API. Examples included in our updated docs: https://docs.gretel.ai/gretel.ai/cli-and-sdk-fundamentals/creating-models#creating-models-with-the-sdk

- When using the SDK, users may download all job artifacts with a new `Model.download_artifacts()` or `RecordHandler.download_artifacts()` method. Examples included in our updated docs: https://docs.gretel.ai/gretel.ai/cli-and-sdk-fundamentals/creating-models#creating-models-with-the-sdk

- Misc bug fixes

Enhancements 🎉
* Add support for "Reference Data" - which allows the passing of misc additional datasets into a Gretel job. These data sets can be used in various ways depending on the context of the job type 33a832ffdcc40ef7a70d1d8f876e90b02e48a476, 899c1eaa4898d1b28eca3119f6c683a619e11096
* Use [`certifi`](https://pypi.org/project/certifi/) CA bundle by default 0ac64dcbfd3dc4bda77921fb60366aa1e8ec4b79
* This default can be overwritten by using `SSL_CERT_FILE` or `REQUESTS_CA_BUNDLE` env variables.
* Allow `--in-data` for manual jobs 1532cb959479fd080744cb126f1cb414f464f5f5
* Add configuration for new model types eee70997fdba91912da8d27f6e0e50b6f730f585
* Add cloud log support for agent based jobs 5d6bb0d949fac69e31034df1c222ae01c8372766

Housekeeping 🧹
* Fix flaky integration test 0fcb2bd81f5386c4cd0ecc09c38124fb4f4f007c
* Increase verbosity of integration test output for easier troubleshooting af483fbec5d674425a048aac82bd17cbd1369426
* Use ECR registry for utility containers 17cb6a1cce544df2ae47a810d79ac740b1ad41e1
* Split up integration tests by component b874e0461d43211320fc55f1ddf3e0690935a842
* Update default python test infrastructure from 3.7 to 3.9 a5e8eb4465d29b65a23cf962221901fce5c22dae

Bugfixes 🐛
* Correctly configure features flag for local jobs b1f38ddf0844dbafecb75337979b6fcd1a466e7b

This RC includes configuration for new model types, which allows training these on a GPU, if one is available.

This RC includes support for "Reference Data" - which allows the passing of misc additional datasets into a Gretel job. These data sets can be used in various ways depending on the context of the job type.

Enhancements 🎉
* Add the ability to run synthetic jobs on CPUs when using `gretel agent` c3ae1bc045b8f13b9e5bce8ea20904e51306e8e7
* Use volume-builder image to prepare data volumes during local runs 143fd41891cb0077fc2882ffe7d8bc72c8a204ad

Bugfixes 🐛
* Model id wasn't being prefixed to agent worker logs 143fd41891cb0077fc2882ffe7d8bc72c8a204ad