SymSpell Spell Check Kotlin

This is a Kotlin Multiplatform implementation of the symspell fuzzy search algorithm. It has been ported from Java implementation of symspell.

Dependency

implementation("com.darkrockstudios:symspellkt:1..0")

Sample

Try out the sample desktop application:

gradlew sampleCompose:run

Try sample here

SymSpell v6.6 (Bigrams)

• the optional bigram dictionary in order to use sentence level context information for selecting best spelling correction.

SymSpell

• The Symmetric Delete spelling correction algorithm reduces the complexity of edit candidate generation and dictionary lookup for a given Damerau-Levenshtein distance.
• It is six orders of magnitude faster (than the standard approach with deletes + transposes + replaces + inserts) and language independent.
• Opposite to other algorithms only deletes are required, no transposes + replaces + inserts. Transposes + replaces + inserts of the input term are transformed into deletes of the dictionary term.
• The speed comes from the inexpensive delete-only edit candidate generation and the pre-calculation.

Customizations

• We replaced the Damerau-Levenshtein implementation with a weighted Damerau-Levenshtein implementation: where each operation (delete, insert, swap, replace) can have different edit weights.
• We added some customizing "hooks" that are used to rerank the top-k results (candidate list). The results are then reordered based on a combined proximity
  • added keyboard-distance to get a dynamic replacement weight (since letters close to each other are more likely to be replaced)
  • do some query normalization before search

Keyboard based Qwerty/Qwertz Distance

There are 2 implementations of the keyboards one is English Qwerty based and other is German Qwertz based implementation we used the adjancey graph of the keyboard for the weights to the connected nodes.

Accuracy Summary

Spellcorrection Strategy	Accuracy	Failures	TP	TN	FP	FN
LUCENE	78.96%	21.04%	5883	481	146	1550
Vanilla SymSpell	88.80%	11.20%	6888	269	358	545
Weighted SymSpell	75.74%	24.26%	5781	324	303	1652
Qwerty Vanilla SymSpell	88.57%	11.43%	6860	279	348	573
Qwerty Weighted SymSpell	75.36%	24.64%	5744	330	297	1689

Example

For 2 terms: 
        slices  
        olives

If the misspelled word is, slives 
both slices and olives is 1 edit distance, 
  so in default case the one with higher frequency will end up in the result.
While with the qwerty based char distance,
 slives is more closer to slices.

The reason for this is in Qwerty Based Keyboard, 
 S and O are too far while V and C are adjacent.

Generation of Deletes

Word deletes are generated with taking edit distance which is minimum of max edit distance and 0.3 * word.length