refactor: remove unnecessary tokenizer options from to_svector functions

[jwnz]

When building the bm25 matrix, we store the tokenizer options. We can reduce some verbosity by referring to these in the bm25_document_to_svector and bm25_query_to_svector functions.

as is:

SELECT bm25_create('documents', 'passage', 'documents_passage_bm25', 'hf', 'google-bert/bert-base-uncased', 0.75, 1.2);

SELECT bm25_document_to_svector('documents_passage_bm25', 'requiring error for due process claim', 'hf', 'google-bert/bert-base-uncased');
SELECT bm25_query_to_svector('documents_passage_bm25', 'requiring error for due process claim', 'hf', 'google-bert/bert-base-uncased');

to be:

SELECT bm25_create('documents', 'passage', 'documents_passage_bm25', 'hf', 'google-bert/bert-base-uncased', 0.75, 1.2);

SELECT bm25_document_to_svector('documents_passage_bm25', 'requiring error for due process claim');
SELECT bm25_query_to_svector('documents_passage_bm25', 'requiring error for due process claim');

[VoVAllen]

@usamoi PTAL