Octue

Summary
Make various user/developer experience improvements to the SDK, mostly to do with cloud storage. These include facilitating more readable and less polluted logging with better object representations.

<!--- SKIP AUTOGENERATED NOTES --->
Contents ([351](https://github.com/octue/octue-sdk-python/pull/351))
**IMPORTANT:** There is 1 breaking change.

Enhancements
- Log starting of service in CLI `start` command
- Give all `Identifiable` subclasses a name-based representation by default, falling back to an ID-based one if no name is provided
- Demote Google Cloud credentials missing environment variable warning to a debug log message
- Get dataset metadata without causing error log messages
- Add `GoogleCloudStorageClient.exists` method
- Raise a more useful error if a cloud bucket can't be found
- Add `cloud_path` property to `Dataset`
- Prefer cloud paths to local paths for datasets in `Manifest.to_cloud`
- **BREAKING CHANGE:** Set default for `Manifest.to_cloud` `store_datasets` parameter to `False`

Fixes
- Allow empty manifests and datasets to pass cloud existence check
- Properly determine datafile paths relative to dataset
- Fall back to the cloud hash or empty hash value if a datafile is not local

Refactoring
- Factor out getting dataset cloud metadata into a method

Testing
- Use latest cloud storage emulator

Other
- Add cortadocodes as an author in `setup.py`

<!--- END AUTOGENERATED NOTES --->

Summary
Simplify instantiation of manifests by specifying their datasets as a dictionary instead of a list with a separate keys parameter.

<!--- SKIP AUTOGENERATED NOTES --->
Contents ([355](https://github.com/octue/octue-sdk-python/pull/355))
**IMPORTANT:** There are 3 breaking changes.

Enhancements
- **BREAKING CHANGE:** Subsume `keys` parameter into `datasets` parameter for `Manifest` instantiation by making `datasets` a dictionary
- **BREAKING CHANGE:** Serialise a manifest's datasets as a dictionary
- **BREAKING CHANGE:** Use latest schema for datasets in manifests

Dependencies
- Use `twined==0.2.0`

Documentation
- Use correct GitHub workflow badge in `README`

<!--- END AUTOGENERATED NOTES --->

Summary
Simplify manifest creation and simplify `octue.resources` cloud storage methods by inferring `project_name` from credentials file.

<!--- SKIP AUTOGENERATED NOTES --->
Contents ([347](https://github.com/octue/octue-sdk-python/pull/347))
**IMPORTANT:** There is 1 breaking change.

Enhancements
- Allow manifests to be defined by a list of dataset paths
- **BREAKING CHANGE:** Infer `project_name` from credentials file and remove its explicit use as a parameter from the following cloud storage methods:
- `GoogleCloudStorageClient`
- `Analysis.finalise`
- `Datafile`
- `Datafile.to_cloud`
- `Dataset.from_cloud`
- `Dataset.to_cloud`
- `Manifest.from_cloud`
- `Manifest.to_cloud`

Refactoring
- Pull cloud path checking into function
- Explicitly include files as a parameter of `Dataset`

Documentation
- Add `manifest` documentation page
- Update version history documentation page
- Fix RST document reference on `logging` page

Style
- Use latest `conventional-commits` pre-commit hook

<!--- END AUTOGENERATED NOTES --->

<!--- SKIP AUTOGENERATED NOTES --->
Contents ([352](https://github.com/octue/octue-sdk-python/pull/352))

Fixes
- Convert non-JSON-serialisable log record args to strings before Pub/Sub log emission

<!--- END AUTOGENERATED NOTES --->

<!--- START AUTOGENERATED NOTES --->
Contents ([345](https://github.com/octue/octue-sdk-python/pull/345))

Enhancements
- Allow `apply_log_handler` to be applied to a logger instance (this is useful for adding a handler to the logger provided by `multiprocessing.get_logger`)

<!--- END AUTOGENERATED NOTES --->

Summary
Update `numpy` in the developer dependencies and template apps to avoid a buffer overflow vulnerability.

<!--- SKIP AUTOGENERATED NOTES --->
Contents ([299](https://github.com/octue/octue-sdk-python/pull/299))

Dependencies
- Increase `numpy` version in developer dependencies and template apps from `1.19.2` to `1.21.0`

Testing
- Decouple monitors tests from cloud storage availability

<!--- END AUTOGENERATED NOTES --->