`remove()` to complement `add()` ? about fuzzysearch HOT 2 CLOSED

So the main complication is that a lot of data is matched by position index in array.

So a temporary removal could be

this.source[idx] = null
this.index[idx] = null
this.store => iterate over all field to remove idx
this.index_map => get unique id of item to remove, then invalidate this using -1, null ?

Then a lot of null check and invalid check.

A more complete removal would require to map all old position to new position.
I'm not sure if this is worth the trouble if rebuilding the index take less than say 200ms
Even if rebuilding the index take like a full second I'd invest in a low priority background worker instead of the complexity of patching the index to new position.

The add() method I can see useful where a client is being drip fed documents.
The remove() I'm not sure of the use case. I suspect it's dependant of the serialization issue.

In any case, I'm not sure the benefit of having this in library vs a soft delete by the user of library.

I'm also working on the assumption that the index can be rebuilt on demand almost instantaneously, I just think it's a waste to recompute it on each search (but many library do pay that cost).
In case of instant search one keystroke = one search.

from fuzzysearch.

As I said right now I gravitate invalidate & rebuild cache.
But I'm interested in your rebuild time.
And if they are unacceptable I'll see if it's worth to properly patch the cache.

from fuzzysearch.