feat(medcat-den): CU-869an5f00 Add remote api

[mart-r]

This is a 2 parter:

1 - new API designed for remote dens

This PR adds optional API that's designed for use with a remote den. The API is not available for local dens.

This API is designed to make all the processing (either fine-tuning or metrics / eval) to happen on the remote.

The imagined workflow would look something like that:

den: Den  # some remote den
model_info = den.list_available_models()[0]  # the info corresponding to model of interest
data = get_data()  # get the data or project IDs
# fine tune
den.finetune_model(model_info, data)
# OR eval
den.evaluate_model(model_info, data)
# the work is done remotely, model is never local to the user

2 - new config options designed for use with remote dens

With remote dens, you generally don't want the user to fine tuning and/or evaluation locally.
Instead, you want this to be done on the remote den so that the experiments can be kept track of.

So this PR adds a few config options / environmental variables (that only affect remote dens):

Environmetnal variable name	Values	Description	Comments
MEDCAT_DEN_REMOTE_ALLOW_PUSH_FINETUNED	bool	Whether to allow locallly fine tuned model to be pushed to remote dens	Defaults to False
MEDCAT_DEN_REMOTE_ALLOW_LOCAL_FINE_TUNE	bool	Whether to allow local fine tuning for remote dens	Defaults to False

So by default, a remote den will not allow local fine-tuning or pushing of a fine-tuned model back to the remote. However, this behaviour can be changed if/when needed. This generally does not affect new models which should still be able to be pushed to the remote den.

[tomolopolis]

Task linked: CU-869an5f00 Move medcat-den to cogstack-nlp