commonMain.net.ormr.fuzzywuzzy.ToStringFunction.kt Maven / Gradle / Ivy

Go to download

Show more of this group Show more artifacts with this name
Show all versions of fuzzy-wuzzy-jvm Show documentation

Kotlin port of the FuzzyWuzzy library

The newest version!

package net.ormr.fuzzywuzzy

/**
 * Transforms and item of type [T] to a [String].
 *
 * @param [T] The type of the item to transform.
 */
public fun interface ToStringFunction {
    /**
     * Transforms the input [item] to a [String].
     *
     * @param [item] the item to transform.
     *
     * @return a string to use for comparing the item.
     */
    public operator fun invoke(item: T): String

    /**
     * A [ToStringFunction] that returns the input string.
     */
    public data object NoProcess : ToStringFunction {
        override fun invoke(item: String): String = item
    }

    public data object StringDefault : ToStringFunction {
        // Kotlin doesn't have a flag for the 'UNICODE_CHARACTER_CLASS' like Java Pattern does,
        // so we need to embed the flag directly into the unicode
        // https://youtrack.jetbrains.com/issue/KT-21094
        private val REGEX: Regex = try {
            """(?U)(?ui)\W""".toRegex()
        } catch (e: IllegalArgumentException) {
            // if parsing the (?U) flag failed, fall back to pattern without it
            """(?ui)\W""".toRegex()
        }

        public override fun invoke(item: String): String = item
            .replaceNonAlphaNumeric(" ")
            .lowercase()
            .trim { it <= ' ' }

        /**
         * Replaces all non-alphanumeric in `this` with [sub].
         */
        private fun String.replaceNonAlphaNumeric(sub: String): String = when {
            REGEX.containsMatchIn(this) -> replace(REGEX, sub)
            else -> this
        }
    }
}