/**
 * A table of conversions (either input to output, or output to input) for
 * a word. Uses the following syntax:
 *
 * ```text
 * # for input-output conversion
 * ICONV <number of entries>
 * ICONV <pattern> <replacement>
 *
 * # for output-input conversion
 * OCONV <number of entries>
 * OCONV <pattern> <replacement>
 * ```
 *
 * Typically, a `pattern` and its `replacement` are just simple strings.
 * This feature is usually used just for normalizing strings containing
 * unusual typographics, like trigraphs and fancy apostrophes.
 *
 * `pattern` does have an unusual undocumented syntax feature: If the `_`
 * underscore character is in the pattern, it may be transformed in one of
 * three ways:
 *
 * - At the start of the `pattern`: Denotes a `RegExp` `^` anchor.
 * - At the end of the `pattern`: Denotes a `RegExp` `$` anchor.
 * - Everywhere else: Entirely ignored.
 *
 * Due to this undocumented feature, patterns are compiled to `RegExp`
 * rather than left as simple strings.
 */
export declare class ConvTable {
    /** The list of patterns and their replacements. */
    table: {
        pattern: RegExp;
        replacement: string;
    }[];
    /**
     * @param pairs - A list of string pair tuples that denotes a pattern and
     *   its replacement.
     */
    constructor(pairs: [string, string][]);
    /**
     * Applies this table's rules against a word and returns the resulting
     * string. Uses the following algorithm, to prevent recursion:
     *
     * - For each position in the word:
     * - ...find any matching rules
     * - ...choose the one that has the longest pattern
     * - ...apply its replacement, and shift the current position past the replacement
     *
     * @param word - The word to transform.
     */
    match(word: string): string;
}