Remove Spacy hard dependency on forte medical ontology

[Piyush13y]

This PR fixes #95.

Description of changes

SpacyProcessor now can be run with the umls_link processor config, without a hard dependency on Forte.
The future scope is to have all medical related ontologies in the Forte-medicine repo, so the default config has been set as so. However, until we reach that point, one can run SpacyProcessor with umls_link config in Forte too, using customized ontology types by passing them as config.

Possible influences of this PR.

No side-effects, ensures consistency and disambiguity in the long run.

[hunterhector]

I think we cannot do setattr here since the list in forte entries are not normal python lists (https://github.com/asyml/forte/blob/master/ftx/onto/clinical.py#L69). This is because we actually need to store reference instead of the actual content. So we should probably do something like getattr(entity, "umls_entities").append(umls_entry)) (Need to try it out to see if it works).

[hunterhector]

i also realize the test case didn't test this part: https://github.com/asyml/forte-wrappers/blob/main/tests/wrappers/spacy_processors_test.py, maybe we should add a small test here?

[Piyush13y]

I have tested the current code, and it does work as expected. I see the umls_entities list populated properly. Maybe because Flist and native lists, both are mutable sequences, we are able to append to Flist through setattr too. I can try using and testing getattr(entity, "umls_entities").append(umls_entry)) as well.

I will add the missing test case for this use case.

[Piyush13y]

Added "umls_link" working to the existing test cases. Other tests run within seconds, but this one goes on for a very long time on my local machine. And I don't think tests are enabled in forte-wrapper's CI pipeline?