Python-pachyderm

- Parity with Pachyderm 2.0.0, including support for Global IDs, and corresponding API changes
- Support for Pachyderm's new spout semantics (added in Pachyderm 1.12)
- Spout-manager has been consolidated into core client, as it's no longer necessary with this change.
- Remove returns from methods whose RPC calls return a negligible protobuf (328)
- `put_files` supports putting individual files
- `Client()` constructor reads connection settings from the Pachyderm config
- API methods that take a commit now accept tuples of the form `(repo, branch, commit_id, repo_type)`
- Support Pachyderm root tokens
- Updated examples

- Support for constructing a `Client` from a Pachyderm config file (PR 220)
- Support for pachyderm v1.11.2 (PR 213)
- Tweaks to debug service functions
- Support for build-step enabled pipelines (PR 213)
- Switched `create_python_pipeline` to use build-step enabled pipelines (PR 213)
- Add support for authenticating with OIDC ID tokens (PR 228)
- Clients automatically authenticate with an ID token in the PACH_PYTHON_OIDC_TOKEN env var (PR 236)
- Add the metadata field to the response from create_pipeline (PR 234)
- Add new_from_config to construct a client from a pachctl config file (PR 220)

- Changes to the interface of enums dynamically generated from protos (PR 207)
- Changes to the `SpoutManager` interface to better support commits (PR 206)
- Added support for putting or deleting many PFS files in an atomic commit (PR 204)
- Deprecated `Client.put_file_bytes` with an iterable of bytestrings (PR 204)
- Added version number to build (PR 203)

- Support for pachyderm v1.11 (PR 201)
- Deprecated `Client.get_admins` and `Client.modify_admins`
- Added support getting/setting cluster role bindings
- Added support for OIDC login
- Added the `full` flag to `Client.inspect_job`
- Added `sidecar_resource_limits` for `Client.list_job`
- Added Loki backend query support for log fetch functions
- Better support for working with pipeline spec files (PR 200)
- Added `parse_json_pipeline_spec` and `parse_dict_pipeline_spec` for parsing JSON spec files
- Added `create_pipeline_from_request` for creating a pipeline from a parsed pipeline spec
- Use `sh` rather than `bash` in python pipelines for better compatibility with custom images (PR 196)

- Allow for customization of spout directory for testing (PR 188)
- Fixed `put_file_bytes` failing for large files (PR 189)

- Support file-like operations on `get_file` results (PR 187)