Forcings Engine Data Provider Base Interface

[program--]

[!IMPORTANT] This PR replaces #720.

This PR adds the ForcingsEngineDataProvider<DataType, SelectionType> class. This class is intended to serve as the base class for Lumped, Gridded, and Unstructured forcings engine data providers. The base class handles timing and updating the underlying BMI instance, so derived classes need only implement the value querying requirements.

Internally, this PR also adds the detail::ForcingsEngineStorage class, which composes a map of initialization file names to Python BMI Adapter shared pointers. The design of this is to allow the Forcings Engine data providers to store a mask of some sort to simplify querying from the forcings engine, i.e. for a Gridded instance, the mask may be a bounding box. This need is due to the implementation of the forcings engine, and should simplify the external requirements for a NGen run using the Python Forcings Engine.

Additionally, this PR includes some minor changes to the forcings library. Namely, the GenericDataProvider class is refactored into a type alias, and the NullForcingProvider implementation is split into header/source files rather than being header-only. The latter change is to ensure the forcings CMake target is always build-able.

Additions

• ForcingsEngineDataProvider: templated base class.
• detail::ForcingsEngineStorage: shared storage between data providers.
• detail::assert_forcings_engine_requirements: helper for ensuring runtime dependencies are available.
• detail::parse_time: helper for parsing timestamp strings.
• constexpr constants for default time format string and forcings engine python information.

Changes

• NullForcingProvider: split implementation into header/source file.
• GenericDataProvider: converted into type alias.

Checklist

• [X] PR has an informative and human-readable title
• [X] Changes are limited to a single goal (no scope creep)
• [X] Code can be automatically merged (no conflicts)
• [X] Code follows project standards (link if applicable)
• [X] Passes all existing automated tests
• [X] Any change in functionality is tested
• [X] New functions are documented (with a description, list of inputs, and expected output)
• [X] Placeholder code is flagged / future todos are captured in comments
• [X] Project documentation has been updated (including the "Unreleased" section of the CHANGELOG)
• [X] Reviewers requested with the Reviewers tool :arrow_right:

[PhilMiller]

The shadowed variable bmi_ is a substantial concern. The other comments are inconsequential.

[PhilMiller]

The main header here can only be included with NGEN_WITH_PYTHON enabled, but I'm not sure that's actually something we need to guard against in the header itself. Implementations and clients are going to need to guard against it anyway. I guess it would save clients conditionalizing both the #include and the usage site.