search from a subset of indexed documents

[roynirmal]

Is it possible to index the entire corpus and yet search within a subset of the documents by passing the corresponding doc_ids?

[xhluca]

Yes! Please use weight_mask in retrieve:

bm25s/bm25s/__init__.py

Lines 629 to 631 in 8b5ff10

	weight_mask : np.ndarray
	A weight mask to filter the documents. If provided, the scores for the masked
	documents will be set to 0 to avoid returning them in the results.

The docstring implies only a binary mask is accepted, but you can use any float mask if you want.

[roynirmal]

Thanks a lot! Very helpful :)