Why?

The current implementation only allows to fuzzy search index-able data structures. However, it does not need random access. It iterates the Source from start to finish -> natural fit for iter.Seq.

In my particular case I have a *list.List as the data source. To fuzzy search that I would have to iterate over the list and copy all elements to a newly allocated slice.

The current implementation only allows to fuzzy search index-able data structures. However, it does not need random access. It iterates the Source from start to finish -> natural fit for iter.Seq.

In my particular case I have a *list.List as the data source. To fuzzy search that I would have to iterate over the list and copy all elements to a newly allocated slice.

Understood and your point is valid but now the API is stable. Does fuzzy.Find(pattern, slices.Collect(seq)) work for you?

Understood and your point is valid but now the API is stable. Does fuzzy.Find(pattern, slices.Collect(seq)) work for you?

Well it would certainly work however I want to avoid unnecessary allocations/iteration but slices.Collect internally just walks the iterator and allocates on the fly.

My proposal does not change the interface, it just adds two new functions.
Though, I moved the matching code in the newly created ones and then call those from the original FindFrom (where I transformed the Source into an iterator) to prevent code duplication.

The API/behavior of the existing functions remains unchanged though.

Understood and your point is valid but now the API is stable. Does fuzzy.Find(pattern, slices.Collect(seq)) work for you?

Well it would certainly work however I want to avoid unnecessary allocations/iteration but slices.Collect internally just walks the iterator and allocates on the fly.

My proposal does not change the interface, it just adds two new functions. Though, I moved the matching code in the newly created ones and then call those from the original FindFrom (where I transformed the Source into an iterator) to prevent code duplication.

The API/behavior of the existing functions remains unchanged though.

Agreed. Merging. Thanks.