Add model metadata

[isaranto]

Add the option of listing available models in the package.
After installing the package the user can run python -m liftwing and get a list of available models along with the API Documentation link and the model card.
When adding a new model the developer needs to add the relevant metadata for the model (create an instance of a pydantic ModelMetadata class) and add that to the init.py of liftwing. Then main.py picks all metadata and prints them.

Alternatively this could be done using a --list argument but I thought that it wasn't needed for now to make things simpler.
If however we plan to add more functionality to this in the future it would be better to do it from now. Let me know what you think.

 python -m liftwing
(
    name=Language-agnostic revert risk,
    classname=RevertRiskAPIModel,
    api_documentation_url=https://api.wikimedia.org/wiki/Lift_Wing_API/Reference/Get_reverted_risk_language_agnostic_prediction,
    wmf_model_card_url=https://meta.wikimedia.org/wiki/Machine_learning_models/Proposed/Language-agnostic_revert_risk
)

[mercelisvaughan]

after reading the instructions this looks good to me