Citrine

This release brings some quality of life improvements to `PredictorEvaluationWorkflow` helper functions, `Dataset` listing, and Documentation

**What’s New**
* Documentation for [`DataSourceDesignSpace`](https://citrineinformatics.github.io/citrine-python/workflows/design_spaces.html#data-source-design-space) has been added

**Improvements**
* Dataset listing performance is improved by increasing the default page size
* Project description has been added to the citrine package on pypi.org
* Type hints has been improved for Lists, Mappers, and Sets. Errors are now raised when setting invalid types.

**Fixes**
* Bugs causing `wait_while_executing` and `get_by_name_or_create` helper functions to fail when given `PredictorEvaluationWorkflow` and `PredictorEvaluationExecution` have been fixed

This is a small release that adds documentation for a newly released feature.

**What’s New**
* Documentation for [`PredictorEvaluationWorkflow`](https://citrineinformatics.github.io/citrine-python/workflows/predictor_evaluation_workflows.html) has been added.

**Fixes**
* A bug in [`wait_while_validating`](https://github.com/CitrineInformatics/citrine-python/blob/04b4fe444e2709878f1c79e83b0fb881c96cbe6a/src/citrine/jobs/waiting.py#L43) that caused timestamps to be formatted incorrectly when issuing status updates has been fixed.

This is a major feature release that includes the new `PredictorEvaluationWorkflow` and `DataSourceDesignSpace`, along with convenience methods for waiting on modules and executions and fetching descriptors from Data Sources.

**What’s New**
* A `PredictorEvaluationWorkflow` has been added. This workflow allows you to evaluate predictor performance using multiple `PredictorEvaluator`s. The first evaluator we're introducing is the `CrossValidationEvaluator`. This evaluator allows you to configure k-fold cross-validation and request desired metrics, e.g. RMSE, NDME, etc. (See `citrine.informatics.predictor_evaluation_metrics` for all the metrics that can be requested.) This workflow is meant to replace the `PerformanceWorkflow`. Instead of returning results as JSON, `PredictorEvaluationWorkflow` results are objects that allow you to fetch results by response, and computed values are now objects that codify the value associated with each requested metric.
* A `DataSourceDesignSpace` has been introduced which allows a Design Space to be created directly from a Data Source.
* New utility functions `wait_while_validating` and `wait_while_executing` allow you to easily wait while a resource is validating or in progress and fetch the result when it's ready.
* Descriptors can be fetched from data source using `DescriptorMethods.descriptors_from_data_source`.

**Improvements**
* The `IngredientQuantityByProcessAndName` variable definition now includes an optional `units` parameter to better support absolute quantities.
* Documentation for TableConfigs and GEM Tables has been updated.

**Deprecations**
* The `PerformanceWorkflow` has been deprecated in favor of the `PredictorEvaluationWorkflow`.
* The `CrossValidationAnalysis` has been deprecated in favor of the `CrossValidationEvaluator`. The latter is used to configure the new `PredictorEvaluationWorkflow`.

v0.70.0 is a major feature release that brings several predictors out of `ALPHA` and adds the `create_or_update` convenience method.

**What’s New**
* The `SimpleMLPredictor`, `ExpressionPredictor`, and `GraphPredictor` are out of `ALPHA`! Their APIs should be relatively stable and they should perform consistently on the platform. Be sure to check the `experimental` flag to check the support level of your specific configuration. (https://github.com/CitrineInformatics/citrine-python/pull/447)
* The `create_or_update` method in the SDK, which makes it easier to avoid creating many copies of the same object when testing scripts. (https://github.com/CitrineInformatics/citrine-python/pull/442)

**Improvements**
* Project listing performance is improved by increasing the default page size (https://github.com/CitrineInformatics/citrine-python/pull/448)

**Removals**
* The `active` field was deprecated by the more functional `archived` functionality. It has been removed from the AI engine modules (https://github.com/CitrineInformatics/citrine-python/pull/437)

v0.66.1 contains a significant new feature: the ability to archive AI Engine modules and workflows.

**What’s New**
* Archiving AI Engine modules and workflows: An archived module is still accessible but is hidden from lists by default. (https://github.com/CitrineInformatics/citrine-python/pull/406)
* Incorporates the most recent release of gemd-python, 0.14.1 (https://github.com/CitrineInformatics/gemd-python/releases/tag/v0.14.1)

**Improvements**
* Better documentation of the names of feature generated by the `GeneralizedMeanPropertyPredictor`. (https://github.com/CitrineInformatics/citrine-python/pull/432)
* Better documentation of which information is removed from the test set when performing cross validation (https://github.com/CitrineInformatics/citrine-python/pull/438)
* Many of the list methods will now automatically deduplicate records that show up on multiple pages when records are being concurrently created. (https://github.com/CitrineInformatics/citrine-python/pull/439)

**Fixes**
* The `add_columns` and `add_all_ingredients` helper methods on `TableConfig` will no longer remove identifiers from `TableConfig` that were read from the platform. (https://github.com/CitrineInformatics/citrine-python/pull/433)

**Deprecations**
* The `filter_by_tags` and `filter_by_name` methods on Data Objects and Templates in data_concepts.py have been deprecated due to their poor performance, and will be removed in a future release. Please use the `list_by_tags` and `list_by_name` methods instead. (https://github.com/CitrineInformatics/citrine-python/pull/436)

v0.61.0 is a feature release that makes it easier to use `Descriptor` by reading them from `AttributeTemplate` and provides more control over cross validation.

**What’s New**
* `PlatformVocabulary` makes it easy to use `AttributeTemplates` as `Descriptors`. For documentation and examples, see the new [descriptors guide](link here). (https://github.com/CitrineInformatics/citrine-python/pull/424)
* `CrossValidationAnalysisConfiguration` now includes an optional `responses` parameter that allows for explicit control over which predictions are validated (https://github.com/CitrineInformatics/citrine-python/pull/429)

**Improvements**
* The design space builders will issue a warning if the resulting space is large enough that it could consume all of your memory (https://github.com/CitrineInformatics/citrine-python/pull/428)

**Changes**
* The type of the `categories` member of `CategoricalDescriptor` has been changed from a list to a set