Gilda

What's Changed
* Add batched grounding endpoint to web service by bgyori in https://github.com/indralab/gilda/pull/100
* Fix error responses from web service

**Full Changelog**: https://github.com/indralab/gilda/compare/0.10.2...0.10.3

What's Changed
* Use legacy UniProt service to generate terms by bgyori in https://github.com/indralab/gilda/pull/99
* Fix `get_models` function to load models before returning if models not loaded yet by bgyori in https://github.com/indralab/gilda/pull/99

**Full Changelog**: https://github.com/indralab/gilda/compare/0.10.1...0.10.2

* Lazy loading for disambiguation models for faster initialization and lower memory footprint in https://github.com/indralab/gilda/pull/96
* Implement optional SQLite back-end for low-memory environment in https://github.com/indralab/gilda/pull/97
* Update grounding resources

What's Changed
API
* Expose all groundings and namespaces, and allow filtering by bgyori in https://github.com/indralab/gilda/pull/89
* Extend grounder API by bgyori in https://github.com/indralab/gilda/pull/91

Integrations
* Further artifacts for Google add-on by bgyori in https://github.com/indralab/gilda/pull/86

Benchmarks
* Update BioID evaluation and add diff tool by cthoyt in https://github.com/indralab/gilda/pull/88
* Split disambiguation evaluation by cthoyt in https://github.com/indralab/gilda/pull/93
* Generalizations to benchmarks by bgyori in https://github.com/indralab/gilda/pull/94

Documentation, tutorials
* Update Gilda documentation and add custom grounder demo notebook by cthoyt in https://github.com/indralab/gilda/pull/92

**Full Changelog**: https://github.com/indralab/gilda/compare/0.8.3...0.9.0

What's Changed
* Add Google Docs add-on by bgyori in https://github.com/indralab/gilda/pull/84
* Add HGNC alias names as synonyms by bgyori in https://github.com/indralab/gilda/pull/85

**Full Changelog**: https://github.com/indralab/gilda/compare/0.8.2...0.8.3

Package and dependencies
- Gilda is now tested with Python 3.6 and 3.10, and support extended to versions 3.6-3.10
- Unit testing framework migrated from nose to pytest

Representation
- Term status flags are updated with the following changes:
- assertion -> curated
- previous -> former_name
- Extended Term object model to represent original source DB/ID for a term (in case a mapping was applied during resource construction) to improve traceability and enable post-processing logic
- Extended ScoredMach object model to represent subsumed terms to improve traceability and enable post-processing logic

Grounding logic
- New logic to recognize and map between Roman and Arabic numerals in entity texts
- Normalize unicode characters to simple ASCII equivalents when looking up matches

Resources
- Gilda resource files now stored as .gz to significantly decrease download size
- Improved UniProt synonym processing fixing many corner cases
- Include Adeft-derived terms with fewer restrictions

REST API and web app
- Significantly extended UI for webapp with more help and documentation
- Extended auto-generated REST API and improved documentation

Documentation
- Add new Jupyter notebook with disambiguation model training workflow