// Copyright 2022 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
import {TestImportManager} from '/common/testing/test_import_manager.js';
import {LocaleInfo} from './locale_info.js';
/**
* EditingUtil provides utility and helper methods for editing-related
* operations.
*/
export class EditingUtil {
/**
* TODO(https://crbug.com/1331351): Add RTL support.
* Returns data needed by inputController.replacePhrase(). Calculates the new
* caret index and number of characters to be deleted. Only operates on the
* text to the left of the text caret. If multiple instances of `deletePhrase`
* are present, this function will operate on the one closest one to the text
* caret.
* @param value The current value of the text field.
* @param deletePhrase The phrase to be deleted.
*/
static getReplacePhraseData(
value: string, caretIndex: number,
deletePhrase: string): {newIndex: number, deleteLength: number}|null {
const leftOfCaret = value.substring(0, caretIndex);
deletePhrase = deletePhrase.trim();
// Find the right-most occurrence of `deletePhrase`. If we're deleting text,
// prefer the RegExps that include a leading/trailing space to preserve
// spacing.
let re;
if (LocaleInfo.considerSpaces()) {
re = EditingUtil.getPhraseRegex_(deletePhrase);
} else {
re = EditingUtil.getPhraseRegexNoWordBoundaries_(deletePhrase);
}
const reWithLeadingSpace =
EditingUtil.getPhraseRegexLeadingSpace_(deletePhrase);
const reWithTrailingSpace =
EditingUtil.getPhraseRegexTrailingSpace_(deletePhrase);
const leadingSpaceResult =
EditingUtil.getIndexFromRegex_(reWithLeadingSpace, leftOfCaret);
const trailingSpaceResult =
EditingUtil.getIndexFromRegex_(reWithTrailingSpace, leftOfCaret);
const noSpacesResult = EditingUtil.getIndexFromRegex_(re, leftOfCaret);
let newIndex = caretIndex;
let deleteLength = 0;
if (leadingSpaceResult !== -1) {
// Delete one extra character to preserve spacing.
newIndex = leadingSpaceResult;
deleteLength = deletePhrase.length + 1;
} else if (trailingSpaceResult !== -1) {
// Delete one extra character to preserve spacing.
newIndex = trailingSpaceResult;
deleteLength = deletePhrase.length + 1;
} else if (noSpacesResult !== -1) {
// Matched with no spacing.
newIndex = noSpacesResult;
deleteLength = deletePhrase.length;
} else {
// No match.
return null;
}
return {newIndex, deleteLength};
}
/**
* TODO(https://crbug.com/1331351): Add RTL support.
* Calculates the new caret index for `inputController.insertBefore`. Only
* operates on the text to the left of the text caret. If multiple instances
* of `beforePhrase` are present, this function will operate on the one
* closest one to the text caret.
* @param value The current value of the text field.
*/
static getInsertBeforeIndex(
value: string, caretIndex: number, beforePhrase: string): number {
const result =
EditingUtil.getReplacePhraseData(value, caretIndex, beforePhrase);
return result ? result.newIndex : -1;
}
/**
* Returns a selection starting at `startPhrase` and ending at `endPhrase`
* (inclusive). The function operates on the text to the left of the text
* caret. If multiple instances of `startPhrase` or `endPhrase` are present,
* the function will use the ones closest to the text caret.
* @param value The current value of the text field.
*/
static selectBetween(
value: string, caretIndex: number, startPhrase: string,
endPhrase: string): {start: number, end: number}|null {
const leftOfCaret = value.substring(0, caretIndex);
startPhrase = startPhrase.trim();
endPhrase = endPhrase.trim();
let startRe;
let endRe;
if (LocaleInfo.considerSpaces()) {
startRe = EditingUtil.getPhraseRegex_(startPhrase);
endRe = EditingUtil.getPhraseRegex_(endPhrase);
} else {
startRe = EditingUtil.getPhraseRegexNoWordBoundaries_(startPhrase);
endRe = EditingUtil.getPhraseRegexNoWordBoundaries_(endPhrase);
}
const start = leftOfCaret.search(startRe);
let end = leftOfCaret.search(endRe);
if (start === -1 || end === -1) {
return null;
}
// Adjust `end` so that we return an inclusive selection.
end += endPhrase.length;
if (start > end) {
return null;
}
return {start, end};
}
/**
* TODO(https://crbug.com/1331351): Add RTL support.
* Returns the start index of the sentence to the right of the caret.
* Indices are relative to `value`. Assumes that sentences are separated by
* punctuation specified in `EditingUtil.END_OF_SENTENCE_REGEX_`. If no next
* sentence can be found, returns `value.length`.
* @param value The current value of the text field.
*/
static navNextSent(value: string, caretIndex: number): number {
const rightOfCaret = value.substring(caretIndex);
const index = rightOfCaret.search(EditingUtil.END_OF_SENTENCE_REGEX_);
if (index === -1) {
return value.length;
}
// `index` should be relative to `value`;
return index + caretIndex + 1;
}
/**
* TODO(https://crbug.com/1331351): Add RTL support.
* Returns the start index of the sentence to the left of the caret. Indices
* are relative to `value`. Assumes that sentences are separated by
* punctuation specified in `EditingUtil.END_OF_SENTENCE_REGEX_`. If no
* previous sentence can be found, returns 0.
* @param value The current value of the text field.
*/
static navPrevSent(value: string, caretIndex: number): number {
let encounteredText = false;
if (caretIndex === value.length) {
--caretIndex;
}
while (caretIndex >= 0) {
const valueAtCaret = value[caretIndex];
if (encounteredText &&
EditingUtil.END_OF_SENTENCE_REGEX_.test(valueAtCaret)) {
// Adjust if there is whitespace immediately to the right of the caret.
return EditingUtil.BEGINS_WITH_WHITESPACE_REGEX_.test(
value[caretIndex + 1]) ?
caretIndex + 1 :
caretIndex;
}
if (!EditingUtil.BEGINS_WITH_WHITESPACE_REGEX_.test(valueAtCaret) &&
!EditingUtil.PUNCTUATION_REGEX_.test(valueAtCaret)) {
encounteredText = true;
}
--caretIndex;
}
return 0;
}
/**
* TODO(https://crbug.com/1331351): Add RTL support.
* This function analyzes the context and adjusts the spacing of `commitText`
* to maintain proper spacing between text.
* @param value The current value of the text field.
*/
static smartSpacing(value: string, caretIndex: number, commitText: string):
string {
// There is currently a bug in SODA (b/213934503) where final speech results
// do not start with a space. This results in a Dictation bug
// (crbug.com/1294050), where final speech results are not separated by a
// space when committed to a text field. This is a temporary workaround
// until the blocking SODA bug can be fixed. Note, a similar strategy
// already exists in Dictation::OnSpeechResult().
if (!value || EditingUtil.BEGINS_WITH_WHITESPACE_REGEX_.test(commitText) ||
EditingUtil.BEGINS_WITH_PUNCTUATION_REGEX_.test(commitText)) {
return commitText;
}
commitText = commitText.trim();
const leftOfCaret = value.substring(0, caretIndex);
const rightOfCaret = value.substring(caretIndex);
if (leftOfCaret &&
!EditingUtil.ENDS_WITH_WHITESPACE_REGEX_.test(leftOfCaret)) {
// If there is no whitespace before the caret index, prepend a space.
commitText = ' ' + commitText;
}
if (rightOfCaret &&
(!EditingUtil.BEGINS_WITH_WHITESPACE_REGEX_.test(rightOfCaret) ||
EditingUtil.BEGINS_WITH_PUNCTUATION_REGEX_.test(rightOfCaret))) {
// If there are no spaces or there is punctuation after the caret index,
// append a space.
commitText = commitText + ' ';
}
return commitText;
}
/**
* TODO(https://crbug.com/1331351): Add RTL support.
* This function analyzes the context and adjusts the capitalization of
* `commitText` as needed. See below for sample input and output: value:
* 'Hello world.' caretIndex: value.length commitText: 'goodnight world'
* return value: 'Goodnight world'
* @param value The current value of the text field.
*/
static smartCapitalization(
value: string, caretIndex: number, commitText: string): string {
if (EditingUtil.BEGINS_WITH_PUNCTUATION_REGEX_.test(commitText)) {
// If `commitText` begins with punctuation, then it's assumed that it's
// already correctly capitalized.
return commitText;
}
if (!value) {
return EditingUtil.capitalize_(commitText);
}
const leftOfCaret = value.substring(0, caretIndex).trim();
if (!leftOfCaret) {
return EditingUtil.capitalize_(commitText);
}
return EditingUtil.ENDS_WITH_END_OF_SENTENCE_REGEX_.test(leftOfCaret) ?
EditingUtil.capitalize_(commitText) :
EditingUtil.lowercase_(commitText);
}
/** Returns a string where the first character is capitalized. */
private static capitalize_(text: string): string {
return text.charAt(0).toUpperCase() + text.substring(1);
}
/** Returns a string where the first character is lowercase. */
private static lowercase_(text: string): string {
return text.charAt(0).toLowerCase() + text.substring(1);
}
/**
* TODO(akihiroota): Break regex construction into smaller chunks.
* Returns a RegExp that matches on the right-most occurrence of a phrase.
* The returned RegExp is case insensitive and requires that `phrase` is
* separated by word boundaries.
*/
private static getPhraseRegex_(phrase: string): RegExp {
return new RegExp(`(\\b${phrase}\\b)(?!.*\\b\\1\\b)`, 'i');
}
/**
* Similar to above, but doesn't include word boundaries. This is useful for
* languages that don't use spaces e.g. Japanese.
*/
private static getPhraseRegexNoWordBoundaries_(phrase: string): RegExp {
return new RegExp(`(${phrase})(?!.*\\1)`, 'i');
}
/** Similar to above, but accounts for a leading space. */
private static getPhraseRegexLeadingSpace_(phrase: string): RegExp {
return new RegExp(`( \\b${phrase}\\b)(?!.*\\b\\1\\b)`, 'i');
}
/** Similar to above, but accounts for a trailing space. */
private static getPhraseRegexTrailingSpace_(phrase: string): RegExp {
return new RegExp(`(\\b${phrase}\\b )(?!.*\\b\\1\\b)`, 'i');
}
private static getIndexFromRegex_(re: RegExp, str: string): number {
const result = re.exec(str);
return result ? result.index : -1;
}
}
export namespace EditingUtil {
/** Includes full-width symbols that are commonly used in Japanese. */
export const END_OF_SENTENCE_REGEX_ = /[;!.?。．？！]/;
/** Similar to above, but looks for a match at the end of a string. */
export const ENDS_WITH_END_OF_SENTENCE_REGEX_ = /[;!.?。．？！]$/;
export const BEGINS_WITH_WHITESPACE_REGEX_ = /^\s/;
export const ENDS_WITH_WHITESPACE_REGEX_ = /\s$/;
export const PUNCTUATION_REGEX_ =
/[-$#"()*;:<>\\\/\{\}\[\]+='~`!@_.,?%。．？！\u2022\u25e6\u25a0]/g;
export const BEGINS_WITH_PUNCTUATION_REGEX_ =
/^[-$#"()*;:<>\\\/\{\}\[\]+='~`!@_.,?%。．？！\u2022\u25e6\u25a0]/;
}
TestImportManager.exportForTesting(EditingUtil);
Codebase Browser
chromium