PSeitz
PSeitz
```rust NgramTokenizer::new(ngram_length, ngram_length, false)?, ``` Usually you have different lenghts for min and max ngrams. You can check what the ngram tokenizer outputs to see which tokens you get (there...
> @PSeitz would it be possible to let tantivy users add custom aggregations? How much work would that be? It would require a big refactor. Currently the requests, intermediate and...
This increased memory consumption per unique term by 4 bytes (12->16 byte for `ExpUnrolledLinkedList`)
`the g* *d wolf` would match `the great bad wolf` (but not `the great abc bad wolf` )
Can you provide some code to reproduce?
I don't see any facets in your example. Can you provide a minimal example with an assertion?
Can you provide a minimal example with an assertion?
It's unclear what the expectation is and which assert fails. Can you add a _minimal_ example, where a document that should be deleted is still there? You can replace `doc_async`...
I think a precomputed `BitSet` should also work well, similar to what I did in [regex_phrase_weight](https://github.com/quickwit-oss/tantivy/blob/main/src/query/phrase_query/regex_phrase_weight.rs#L101)
> > I think a precomputed `BitSet` should also work well, similar to what I did in [regex_phrase_weight](https://github.com/quickwit-oss/tantivy/blob/main/src/query/phrase_query/regex_phrase_weight.rs#L101) > > Hm, interesting. > > That could be implemented using either...