Closed Bluebie closed 2 years ago
Taking advantage of the new index format which provides words lists to the browser, if a query fails to find word vectors for the word tokens in the query, fastest-levenshtein
is used to find similar words in the dataset and suggest an autocorrection.
Feedback from Dyslexic users has been positive. Deployed to production.
Dyslexic users sometimes enter obscure mispellings like "interpeter" unknowingly, and can face a "no results found" response as a result. Find Sign should be able to do something in this case, to provide a reasonable result or suggestions for spelling correction.
Some options: