Update unit tests + contribution guidelines to support HFhub submissions

[hakunanatasha]

Supercedes #850

TLDR:

• keeps the biodatasets and test_bigbio script
• adapts the test_bigbio_hub script to use the --test_local flag in order to test hub submissions
• fixes minor broken hyperlink in README for contributions guideline
• adapt contribution guidelines for new hub-forward scripting
• includes new unit test that checks for language + license metadata to be explicitly standardized by the values in languages.json and licenses.json

How to run:

1) There is a test folder here: bigbio/hub/hub_repos/test_scitail. This includes a dataloading script and a copy of the bigbiohub.

2) You can test the unit tests in the main bigbio directory (parent of the tests folder) by running the following command:

python -m tests.test_bigbio_hub test_scitail --test_local

This will pass, but you can test the metadata works as intended by modifying the _LANGUAGES and _LICENSES category. I made 2 examples

NOTES

• I found that the default test script has a weird license? APACHE_p2.0; we should look into this?

  • test_metadata now explicitly checks the elements of _LANGUAGES and _LICENSES for the content; this explicitly reads languages.json and licenses.json to ensure that all metadata elements adhere to our curated list. Currently, this is a AssertionError, we may want to think about making this flexible, because a test will fail unless we have already seen this license + language before.

[galtay]

@hakunanatasha I added a few small changes

• import lang_keys and license_keys from modules
• update to use the short license names going to merge now. fan-freaking-tastic work!