Closed bocse closed 2 years ago
I added the following endpoints for KNN searches: /codex/knn/editdistance
to search by Levenshtein, Hamming or LCS distance, and codex/knn/ngram
to search by n-gram similarity or n-gram positional similarity. Implementation details can be found in https://github.com/RaduADumitru/codex/blob/main/codex/src/main/java/org/example/codex/controller/KnnController.java
What is the endpoint I can call to compute k-nearest neighbours between a given string (”anaaremere”) and the words in the dictionarym, while also specifying the distance (Euclid, Hann, Hamming, Manhattan, Levenshtein) ?
Results should be ordered descending by distance (nearest elements first, farthest elements last) and it should also have some parameter for setting a LIMIT (cut-off) (eg. show only first 100 matches)
I looked here https://github.com/bocse/codex/blob/main/codex/src/main/java/org/example/codex/controller/LexemeController.java
Please advise :)