Regex cache (#4858) feat: optimize regex extension performance - Implemented caching mechanism for regex scripts to improve message rendering speed. - Added automatic cache invalidation on script updates. - Optimized regex compilation to reduce CPU usage. --------- Co-authored-by: GhostXia <33112711+GhostXia@users.noreply.github.com>
Signed| @@ -35,6 +35,61 @@ export const SCRIPT_TYPE_UNKNOWN = -1; | ||
| 35 | 35 | const DEFAULT_GET_REGEX_SCRIPTS_OPTIONS = Object.freeze({ allowedOnly: false }); |
| 36 | 36 | |
| 37 | 37 | /** |
| 38 | + * Manages the compiled regex cache with LRU eviction. | |
| 39 | + */ | |
| 40 | +export class RegexProvider { | |
| 41 | + /** @type {Map<string, RegExp>} */ | |
| 42 | + #cache = new Map(); | |
| 43 | + /** @type {number} */ | |
| 44 | + #maxSize = 1000; | |
| 45 | + | |
| 46 | + static instance = new RegexProvider(); | |
| 47 | + | |
| 48 | + /** | |
| 49 | + * Gets a regex instance by its string representation. | |
| 50 | + * @param {string} regexString The regex string to retrieve | |
| 51 | + * @returns {RegExp?} Compiled regex or null if invalid | |
| 52 | + */ | |
| 53 | + get(regexString) { | |
| 54 | + const isCached = this.#cache.has(regexString); | |
| 55 | + const regex = isCached | |
| 56 | + ? this.#cache.get(regexString) | |
| 57 | + : regexFromString(regexString); | |
| 58 | + | |
| 59 | + if (!regex) { | |
| 60 | + return null; | |
| 61 | + } | |
| 62 | + | |
| 63 | + if (isCached) { | |
| 64 | + // LRU: Move to end by re-inserting | |
| 65 | + this.#cache.delete(regexString); | |
| 66 | + this.#cache.set(regexString, regex); | |
| 67 | + } else { | |
| 68 | + // Evict oldest if at capacity | |
| 69 | + if (this.#cache.size >= this.#maxSize) { | |
| 70 | + const firstKey = this.#cache.keys().next().value; | |
| 71 | + this.#cache.delete(firstKey); | |
| 72 | + } | |
| 73 | + this.#cache.set(regexString, regex); | |
| 74 | + } | |
| 75 | + | |
| 76 | + // Reset lastIndex for global/sticky regexes | |
| 77 | + if (regex.global || regex.sticky) { | |
| 78 | + regex.lastIndex = 0; | |
| 79 | + } | |
| 80 | + | |
| 81 | + return regex; | |
| 82 | + } | |
| 83 | + | |
| 84 | + /** | |
| 85 | + * Clears the entire cache. | |
| 86 | + */ | |
| 87 | + clear() { | |
| 88 | + this.#cache.clear(); | |
| 89 | + } | |
| 90 | +} | |
| 91 | + | |
| 92 | +/** | |
| 38 | 93 | * Retrieves the list of regex scripts by combining the scripts from the extension settings and the character data |
| 39 | 94 | * |
| 40 | 95 | * @param {GetRegexScriptsOptions} options Options for retrieving the regex scripts |
| @@ -353,7 +408,7 @@ export function runRegexScript(regexScript, rawString, { characterOverride } = { | ||
| 353 | 408 | } |
| 354 | 409 | }; |
| 355 | 410 | const regexString = getRegexString(); |
| 356 | 411 | const findRegex = regexFromStringRegexProvider.instance.get(regexString); |
| 357 | 412 | |
| 358 | 413 | // The user skill issued. Return with nothing. |
| 359 | 414 | if (!findRegex) { |
| @@ -8,7 +8,7 @@ import { commonEnumProviders, enumIcons } from '../../slash-commands/SlashComman | ||
| 8 | 8 | import { SlashCommandEnumValue, enumTypes } from '../../slash-commands/SlashCommandEnumValue.js'; |
| 9 | 9 | import { SlashCommandParser } from '../../slash-commands/SlashCommandParser.js'; |
| 10 | 10 | import { download, equalsIgnoreCaseAndAccents, escapeHtml, getFileText, getSortableDelay, isFalseBoolean, isTrueBoolean, regexFromString, setInfoBlock, uuidv4 } from '../../utils.js'; |
| 11 | 11 | import { allowPresetScripts, allowScopedScripts, disallowPresetScripts, disallowScopedScripts, getCurrentPresetAPI, getCurrentPresetName, getRegexScripts, getScriptsByType, isPresetScriptsAllowed, isScopedScriptsAllowed, regex_placement, RegexProvider, runRegexScript, saveScriptsByType, SCRIPT_TYPE_UNKNOWN, SCRIPT_TYPES, substitute_find_regex } from './engine.js'; |
| 12 | 12 | import { t } from '../../i18n.js'; |
| 13 | 13 | import { accountStorage } from '../../util/AccountStorage.js'; |
| 14 | 14 | import { getPresetManager } from '../../preset-manager.js'; |
| @@ -1628,6 +1628,8 @@ async function checkCharEmbeddedRegexScripts() { | ||
| 1628 | 1628 | } |
| 1629 | 1629 | } |
| 1630 | 1630 | |
| 1631 | + // Clear cache and reload scripts | |
| 1632 | + RegexProvider.instance.clear(); | |
| 1631 | 1633 | await loadRegexScripts(); |
| 1632 | 1634 | } |
| 1633 | 1635 | |