Curate additional DOID-MESH mappings and contribute to DOID

[bgyori]

This PR does the following:

• Add new DOID-MESH predictions based on a new script that predicts largely correct, one-to-one mappings
• Curates these new mappings to find 152 new correct mappings
• Adds a script to add all non-redundant, curated DOID-MESH mappings (i.e., not just new ones but also ones that were curated earlier) to the editable OWL version of DOID in version control

Note that some, relatively less reliable predictions originally produced by https://github.com/biopragmatics/biomappings/blob/master/scripts/generate_doid_mappings.py were removed, though these do contain at least some correct predictions, though in principle, this new script would have found the same matches if they were non-redundant and one-to-one.

[bgyori]

@cthoyt I realize we haven't wrapped this up yet. Currently it's stuck due to having to decide on the conventions for MONDO IDs, which is actually dependent on how we decide to handle MONDO on the Bioregistry side. To recap, choices are:

• (a) MONDO has namespace-embedded-in-LUI since it's in OBOFoundry
• (b) MONDO doesn't have namespace-embedded-in-LUI since identifiers.org doesn't explicitly say that it has

[cthoyt]

I’d like to go with B. We shouldn’t make any assumptions about what identifiers.org might do for resources it hasn’t curated.