typo-detector

Typo detector library in Scala.

Typical use case: certain words need to be found in a text, however they might be misspelled in a variety of ways. Instead of trying to hardcode the possible misspellings of such an input typo-detector might be used instead.

Usage

Add dependency

Maven

<dependency>
  <groupId>io.github.antivanov</groupId>
  <artifactId>typo-detector_3</artifactId>
  <version>0.4.0</version>
</dependency>

SBT

For Scala 3:

libraryDependencies += "io.github.antivanov" %% "typo-detector" % "0.4.0"

For Scala 2:

libraryDependencies += ("io.github.antivanov" %% "typo-detector" % "0.4.0")
  .cross(CrossVersion.for2_13Use3)

Make sure you have -Ytasty-reader compiler flag set:

ThisBuild / scalacOptions := Seq(
  "-Ytasty-reader",
)

Code examples

Searching for (maybe mistyped) words in a text

Enhancing the String class with containsExactOrTypoOf method

import io.github.antivanov.typo.detector.syntax._

"Quick bron fox jumps over the lazy dog".containsExactOrTypoOf("brown fox")

Using API object TypoDetector if using implicits is not desirable

import io.github.antivanov.typo.detector.TypoDetector

TypoDetector.containsExactOrTypoOf("Quick bron fox jumps over the lazy dog", "brown fox")

Detecting if a string is a typo of another string

Enhancing the String class with isTypoOf method

import io.github.antivanov.typo.detector.syntax._

val acknowlegmentText = "acknowlege"
val isAcknowledged = acknowlegmentText.isTypoOf("acknowledge")

Using API object TypoDetector if using implicits is not desirable

import io.github.antivanov.typo.detector.TypoDetector

val acknowlegmentText = "acknowlege"
val isAcknowledged = TypoDetector.isTypoOf(acknowlegmentText, "acknowledge")

Detecting if another string is equal to current string or is its typo

equalsOrTypoOf is very similar in usage to isTypoOf, but returns true if another string is either a typo or is equal to the current string

Specifying how many wrong symbols is considered a typo

By default the number of wrongly typed symbols for a string to be considered a typo is 2, in the case of longer strings it might make sense to allow for more mistyped symbols to detect typos, for example:

Enhancing the String class with isTypoOf method

import io.github.antivanov.typo.detector.syntax._

val misspelledWord = "gementeradverkiezingen"
val isMisspelled = misspelledWord.isTypoOf("gemeenteraadsverkiezingen", maxMistypedSymbols = 5)

Using API object TypoDetector if using implicits is not desirable

import io.github.antivanov.typo.detector.TypoDetector

val misspelledWord = "gementeradverkiezingen"
val isMisspelled = TypoDetector.isTypoOf(misspelledWord, "gemeenteraadsverkiezingen", maxMistypedSymbols = 5)

Computing the edit distance between the strings

editDistanceFrom computes the Levenshtein distance between the strings

Enhancing the String class with editDistanceFrom method

import io.github.antivanov.typo.detector.syntax._

val acknowlegmentText = "acknowlege"
val distance = acknowlegmentText.editDistanceFrom("acknowledge")

Using API object TypoDetector if using implicits is not desirable

import io.github.antivanov.typo.detector.TypoDetector

val acknowlegmentText = "acknowlege"
val distance = TypoDetector.editDistanceFrom(acknowlegmentText, "acknowledge")

How it works

Under the hood the library computes the Levenshtein distance between strings using the Wagner-Fischer algorithm. If the distance is lower than a certain number, the two strings are considered to be a typo.