Performance improvements for collection-level queries
What's Changed
The main change is added parallelization, within this package, for queries over `SOMACollection` objects. Additional and significant performance improvements for queries on each single SOMA within a given collection are recently obtained via the `TileDB-Py` 0.17.0 release which includes TileDB core 2.11.0.
Major PRs:
* Some thread-pool-executor parallelizations, part 1/n by johnkerl in https://github.com/single-cell-data/TileDB-SingleCell/pull/238
* Parallelize soma collection queries by Shelnutt2 in https://github.com/single-cell-data/TileDB-SingleCell/pull/239
* Use max_workers config for bounded thread-pool execution by johnkerl in https://github.com/single-cell-data/TileDB-SingleCell/pull/240
* Parallelize constructors in collection-iterator by johnkerl in https://github.com/single-cell-data/TileDB-SingleCell/pull/241
* SOMA-list query option, and further thread-pool experiments by johnkerl in https://github.com/single-cell-data/TileDB-SingleCell/pull/242
* Wording updates for doc cover page by johnkerl in https://github.com/single-cell-data/TileDB-SingleCell/pull/236
Minor/internal PRs:
* Minor tweak to if-exists logic by johnkerl in https://github.com/single-cell-data/TileDB-SingleCell/pull/231
* Do not throw on "in" check on unpopulated collection by johnkerl in https://github.com/single-cell-data/TileDB-SingleCell/pull/233
* Use macos-12 for CI due to GitHub brownout of macos-10.15 by johnkerl in https://github.com/single-cell-data/TileDB-SingleCell/pull/234
**Full Changelog**: https://github.com/single-cell-data/TileDB-SingleCell/compare/0.1.4...0.1.5