Update README.md

Add type hint for `texts` argument in `tokenize` function (#44)

* Use `time.monotonic` instead of `time.time`

* Add type hint for `texts` argument in `tokenize` function

Update numba backend to use numpy sorting instead of internal sorting

Change to np.zero instead of np.empty

Add _compute_relevance_from_scores_jit_ready as an alternative

refactor moving bm25s to compute relevance from eager score as a stat…

…ic method rather than a function

Update hf.py

Update hf.py

Bug fix and add link (#28)

* Update HF readme template

* Fix type for def retrieve

* Add tests for retrieve function

* Update retrieve to take tuples

Update __init__.py