Citrine

This is a maintenance release for a bug-fix when reading GemTables.

Fixes
* Correctly read tables specified as (uid, version) pairs (https://github.com/CitrineInformatics/citrine-python/pull/678)

This is a maintenance release that includes improvements to convenience methods like `Dataset.delete_contents` and `TableConfig.add_all_ingredients`, along with several small fixes and documentation changes.

Improvements
* Added an optional "Are you sure?" prompt to the `Dataset.delete_contents` method (https://github.com/CitrineInformatics/citrine-python/pull/669)
* Added label variables and columns to the output of the `TableConfig.add_all_ingredients` convenience method (https://github.com/CitrineInformatics/citrine-python/pull/661)
* Describe how regression and classification problems are setup with `AutoMLPredictor` in the documentation (https://github.com/CitrineInformatics/citrine-python/pull/675)

Fixes
* Raised a more informative error when trying to get a `TableConfig` without a `uid` (https://github.com/CitrineInformatics/citrine-python/pull/671)
* Added missing imports to code snippets in documentation (https://github.com/CitrineInformatics/citrine-python/pull/672)
* Updated deprecated function calls in documentation (https://github.com/CitrineInformatics/citrine-python/pull/673)

Improvements
* We added a tool that scans the repository for outdated annotations and other deprecated aspects of the API. It raises a flag when it finds an issue. 654
* Design workflow executions that were taking a long time would time out and fail with an opaque error message. The error message now includes the execution's unique ID for easier debugging. 666
* We added `read_to_memory` to the GEMTables interface. This method allows you to download a table (formatted as a CSV in a string) into a variable rather than requiring a round-trip to disk. 667

Fixes
* Previously, the `pandas<=1.1` and `numpy>=1.2.0` combo would cause numpy to throw `dtype` alias deprecation warnings. We changed the pandas dependency requirement to `pandas>=1.1.0` to avoid these warnings. 659
* The documentation on predictors used out-of-date argument names when illustrating how to use chemical formula and molecular structure featurizers . They have been updated. 660
* There were a few private methods whose names did not have a leading underscore. We fixed that. 663
* Added a setter method to fix a minor issue regarding the `TableConfig` class's `uid` attribute. 664

What's New
* A new builder module for auto-configuring platform assets. Users can create an `AutoConfigureWorkflow`, which configures default assets starting from a material, GEM table, or predictor. 653

Improvements
* An alias for accessing the Citrine-scope UID on all data concepts objects by calling `object.uid`. 658
* An alias on Table Configs for accessing the field `config_uid` by simply calling `config.uid`. 655

Fixes
* Warnings are by default enabled for deprecated methods. 656

This release includes the addition of a new collection `GEMDResourceCollection` for accessing generic data objects from the Citrine Platform. Users can now get, list, register, update, and delete GEMD objects, regardless of their type, through methods on this generic collection accessed by `project.gemd` or `dataset.gemd`.

What's New
* Client support for interacting with generic data objects on the Citrine Platform. 646

What's New
* Defining the components of Table Configs can now be done using Templates or IDs instead of just LinkByUIDs. This matches the access patterns of other elements of the citrine-python API and should be more convenient for using a collection of GEMD Objects to build exactly the table you need. 648

Improvements
* We now handle URLs with a wider range of characters for scopes and identifiers. We now have full support for all characters allowed by the [GEMD specification](https://citrineinformatics.github.io/gemd-docs/specification/unique-identifiers/#alternative-ids). 650

Fixes
* Plus some documentation and test fixes 647 649