perf: improve alignment::distance::levenshtein and alignment::distance::bounded_levenshtein on strrings where distance is small #522
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Probably, it is better to use the implementation of the Ukkonen algorithm from
editdistancek
.Before:
After:
The implementation works in time O(nd) where d is the edit distance between strings (+simd optimization). The implementation outperforms the previous version except when the edit distance is close to the maximum possible value. You can check more benchmarks at https://github.com/nkkarpov/editdistancek . On long random strings (1k) with random edits the version from
editdistancek
can achieve a 24x speedup compared totriple_accell
when the percent of edits is less than 1. When the percent of edits is at most 20, and the upper bound is equal to 2 times the real distance, the new implementation has at least a 2.5x speedup. The algorithm is threshold oblivious, and the upper bound on the distance is used only for memory allocation (the algorithm requires O(k) auxiliary memory where k is the upper bound on the distance).