Marcus Pousette

Main issue from the logs is ` failed to find default addresses `

Will try to get a PR up soon as I have time

@tmpfs I have not given up on this task, but I have been very busy this autumn and have not have time to work on this lib at all. Hopefully...

Not at the moment. I will do some work on this lib in the near future (weeks, month time) to improve this library, but have not considered this yet since...

Sounds great! I have also had the idea of adding this at some point. But never had the time to dig into the details of common/efficient ways of implementing it....

You can make the index case-insensitive by using a "to lower case" ```Filter``` that you would pass both when adding documents, but also when querying. Great to hear that! 👍

I like that idea. We want to the API the be very generic so that users could repurpose however they want. The only thing I am thinking about it is...

Another note, I have already made the "score calculation" in to a [trait](https://github.com/quantleaf/probly-search/blob/master/src/query/score/calculator.rs) see [bm25](https://github.com/quantleaf/probly-search/blob/master/src/query/score/default/bm25.rs) and [zero_to_one](https://github.com/quantleaf/probly-search/blob/master/src/query/score/default/zero_to_one.rs) for implementations of this trait. Maybe this could come handy in some way,...

It might be necessary yes, try it out! One thing to test also (as I previously mentioned). Is to implement an N-gram tokenizer. This should be a very simple thing...

Great work. Looks like a very clean solution. That sounds very convenient if it is behind a feature flag! Yes. I see! I was playing around with the segmentation a...