I noticed the project was based on typo-js, but the API you provide is really better.
So thank you.
When looking at typo-js, I noticed a couple of pull requests that were not merged.
A very interesting one is cfinke/Typo.js#28.
By refactoring the edits1 function, the commiter claims 10x improvements.
I decided to create a simple perf test to see if it was true.
Results:
Before edits1 modification:
231 suggestions in 77047 ms
After:
231 suggestions in 9215 ms
So that's an 8-fold improvement over the original version, without losing any suggestion.
Credit goes to @tonioloewald for his amazing work.
Another commiter claims that we can get much better performances if we do not use distance2.
According to my bench, that's indeed another 20-fold improvement but I need to run some more tests as it significantly reduces the number of suggestions.
So I'll keep you posted on my progress but this one is totally safe to merge and it really improves performance.
Hello there!
I noticed the project was based on typo-js, but the API you provide is really better. So thank you.
When looking at typo-js, I noticed a couple of pull requests that were not merged. A very interesting one is cfinke/Typo.js#28.
By refactoring the edits1 function, the commiter claims 10x improvements.
I decided to create a simple perf test to see if it was true.
Results:
So that's an 8-fold improvement over the original version, without losing any suggestion. Credit goes to @tonioloewald for his amazing work.
Another commiter claims that we can get much better performances if we do not use distance2. According to my bench, that's indeed another 20-fold improvement but I need to run some more tests as it significantly reduces the number of suggestions.
So I'll keep you posted on my progress but this one is totally safe to merge and it really improves performance.