The ultimate URL purifier
此脚本不应直接安装。它是供其他脚本使用的外部库，要使用该库请加入元指令 // @require https://update.cn-greasyfork.org/scripts/492078/1499254/pURLfy.js
class Purlfy extends EventTarget {
// Static properties
/**
* Returns the version of the library.
* @returns {string} The version of the library.
*/
static get version() {
return "0.3.11";
};
/**
* A TextDecoder object used internally.
* @type {TextDecoder}
*/
static #decoder = new TextDecoder();
/**
* A DOMParser object used internally.
* @type {DOMParser | null}
*/
static #domParser = typeof DOMParser !== "undefined" ? new DOMParser() : null;
/**
* The constructor of the AsyncFunction class.
* @type {Function}
*/
static #AsyncFunction = async function () { }.constructor;
/**
* The initial statistics object. (All values are 0)
* @type {Object}
*/
static #zeroStatistics = {
url: 0,
param: 0,
decoded: 0,
redirected: 0,
visited: 0,
char: 0
};
/**
* The default acts for URL purification.
* @type {Object}
*/
static #acts = {
url: decodeURIComponent,
base64: s => { // https://developer.mozilla.org/en-US/docs/Web/API/Window/btoa#unicode_strings
s = s.replaceAll('_', '/').replaceAll('-', '+');
const bytes = Uint8Array.from(atob(s), (m) => m.codePointAt(0));
return Purlfy.#decoder.decode(bytes);
},
slice: (s, startEnd) => {
const [start, end] = startEnd.split(":");
return s.slice(parseInt(start), end ? parseInt(end) : undefined)
},
regex: (s, regex) => {
const r = new RegExp(regex);
const m = s.match(r);
return m ? m[0] : "";
},
dom: (s) => Purlfy.#domParser.parseFromString(s, "text/html"),
sel: (s, selector) => s.querySelector(selector),
attr: (e, attr) => e.getAttribute(attr),
text: (e) => e.textContent,
};
// Instance properties
/**
* Whether to enable the fetch mode.
* @type {boolean}
*/
fetchEnabled = false;
/**
* Whether to enable the lambda mode.
* @type {boolean}
*/
lambdaEnabled = false;
/**
* The maximum number of iterations for purification.
* @type {number}
*/
maxIterations = 5;
/**
* The logger function.
* @type {Function}
*/
#log = console.log.bind(console, "\x1b[38;2;220;20;60m[pURLfy]\x1b[0m");
/**
* The fetch function.
* @type {Function}
*/
#fetch = fetch.bind(globalThis);
/**
* The statistics object.
* @type {Object}
*/
#statistics = { ...Purlfy.#zeroStatistics };
/**
* The rules object.
* @type {Object}
*/
#rules = {};
/**
* Creates a new instance of the Purlfy class.
* @param {Object} [options] The options for the instance.
* @param {boolean} [options.fetchEnabled] Whether to enable the fetch mode.
* @param {boolean} [options.lambdaEnabled] Whether to enable the lambda mode.
* @param {number} [options.maxIterations] The maximum number of iterations for purification.
* @param {Object} [options.statistics] The statistics object.
* @param {Function} [options.log] The logger function.
* @param {Function} [options.fetch] The fetch function.
*/
constructor(options) {
super();
this.fetchEnabled = options?.fetchEnabled ?? this.fetchEnabled;
this.lambdaEnabled = options?.lambdaEnabled ?? this.lambdaEnabled;
this.maxIterations = options?.maxIterations ?? this.maxIterations;
Object.assign(this.#statistics, options?.statistics);
this.#log = options?.log ?? this.#log;
this.#fetch = options?.fetch ?? this.#fetch;
}
// Static methods
/**
* Checks if the given value is of the given type or undefined.
* @param {*} value The value to check.
* @param {string} type The type to check.
* @returns {boolean} Whether the given value is of the given type or undefined.
*/
static #udfOrType(value, type) {
return value === undefined || typeof value === type;
}
/**
* Checks if the given URL object's search string follows the standard format.
* @param {URL} urlObj The URL object to check.
* @returns {boolean} Whether the given URL object's search string follows the standard format.
*/
static #isStandard(urlObj) {
return urlObj.searchParams.toString() === urlObj.search.slice(1);
}
/**
* Checks if the given item is an object.
* @param {*} item The item to check.
* @returns {boolean} Whether the given item is an object.
* @see https://stackoverflow.com/questions/27936772
*/
static #isObject(item) {
return (item && typeof item === 'object' && !Array.isArray(item));
}
/**
* Merges the given objects deeply.
* @param {Object} target The target object to merge into.
* @param {...Object} sources The source objects to merge.
* @returns {Object} The merged object.
* @see https://stackoverflow.com/questions/27936772
*/
static #mergeDeep(target, ...sources) { // TODO: handle rules conflict (e.g. "path" and "path/")
if (!sources.length) return target;
const source = sources.shift();
if (Purlfy.#isObject(target) && Purlfy.#isObject(source)) {
for (const key in source) {
if (Purlfy.#isObject(source[key])) {
if (!target[key]) Object.assign(target, { [key]: {} });
Purlfy.#mergeDeep(target[key], source[key]);
} else {
Object.assign(target, { [key]: source[key] });
}
}
}
return Purlfy.#mergeDeep(target, ...sources);
}
/**
* Applies the given acts to the given input.
* @param {string} input The input to apply the acts to.
* @param {string[]} acts The acts to apply.
* @param {Function} logFunc The logger function.
* @returns {string | null} The result of applying the given acts to the given input.
*/
static #applyActs(input, acts, logFunc) {
let dest = input;
for (const cmd of (acts)) {
const name = cmd.split(":")[0];
const arg = cmd.slice(name.length + 1);
const act = Purlfy.#acts[name];
if (!act) {
logFunc("Invalid act:", cmd);
dest = null;
break;
}
try {
dest = act(dest, arg);
} catch (e) {
logFunc(`Error processing input with act "${name}":`, e);
dest = null;
break;
}
}
if (typeof dest === "string" || dest instanceof URL) {
return dest.toString();
} else {
return null;
}
}
// Instance methods
/**
* Clears the statistics.
* @returns {void}
*/
clearStatistics() {
const increment = {};
for (const [key, value] of Object.entries(this.#statistics)) {
increment[key] = -value;
}
this.#incrementStatistics(increment);
}
/**
* Clears the rules.
* @returns {void}
*/
clearRules() {
this.#rules = {};
}
/**
* Gets the statistics.
* @returns {Object} The statistics.
*/
getStatistics() {
return { ...this.#statistics };
}
/**
* Imports the given rules.
* @param {...Object} rulesets The rulesets to import.
* @returns {void}
*/
importRules(...rulesets) {
Purlfy.#mergeDeep(this.#rules, ...rulesets);
}
/**
* Checks if the given rule is valid.
* @param {Object} rule The rule to check.
* @returns {boolean} Whether the given rule is valid.
*/
#validRule(rule) {
if (!rule || !rule.mode || !rule.description || !rule.author) return false;
if ((rule.acts ?? []).includes("dom") && !Purlfy.#domParser) return false; // Feature detection for DOMParser
switch (rule.mode) {
case "white":
return Array.isArray(rule.params);
case "black":
return Array.isArray(rule.params) && Purlfy.#udfOrType(rule.std, "boolean");
case "param":
return Array.isArray(rule.params) && (rule.acts === undefined || Array.isArray(rule.acts)) && Purlfy.#udfOrType(rule.continue, "boolean");
case "regex":
return Array.isArray(rule.regex) && Array.isArray(rule.replace) && Purlfy.#udfOrType(rule.continue, "boolean") && rule.regex.length === rule.replace.length;
case "redirect":
return this.fetchEnabled && Purlfy.#udfOrType(rule.ua, "string") && Purlfy.#udfOrType(rule.headers, "object") && Purlfy.#udfOrType(rule.continue, "boolean");
case "visit":
return this.fetchEnabled && Purlfy.#udfOrType(rule.ua, "string") && Purlfy.#udfOrType(rule.headers, "object") && (rule.acts === undefined || Array.isArray(rule.acts)) && Purlfy.#udfOrType(rule.continue, "boolean");
case "lambda":
return this.lambdaEnabled && (typeof rule.lambda === "string" || rule.lambda instanceof Purlfy.#AsyncFunction) && Purlfy.#udfOrType(rule.continue, "boolean");
default:
return false;
}
}
/**
* Iteratively matches the longest rule for the given URL parts.
* @param {string[]} parts The URL parts to match.
* @returns {Object|null} The matched rule.
*/
#matchRule(parts) {
let fallbackRule = null; // Most precise fallback rule
let currentRules = this.#rules;
for (const part of parts) {
if (currentRules.hasOwnProperty("")) {
fallbackRule = currentRules[""];
}
if (currentRules.hasOwnProperty(part + "/")) {
currentRules = currentRules[part + "/"]; // Exact match - continue to the next level
} else if (currentRules.hasOwnProperty(part)) {
const rule = currentRules[part];
if (this.#validRule(rule)) {
return rule; // Exact match found
}
} else { // No exact match found, try to match with regex
let found = false;
// Iterate through current rules to match RegExp
for (const [key, val] of Object.entries(currentRules)) {
if (!key.startsWith("/")) continue; // Skip non-RegExp keys
try {
const sub = key.endsWith("/"); // Has sub-rules
const regexStr = sub ? key.slice(1, -1) : key.slice(1);
if (regexStr === "") continue; // Skip empty regex
const regex = new RegExp(regexStr);
if (regex.test(part)) { // Regex matches
if (!sub && this.#validRule(val)) {
return val; // Regex match found
} else if (sub) {
currentRules = val; // Continue to the next level
found = true;
break;
}
}
} catch (e) {
this.#log("Invalid regex:", key.slice(1));
}
}
if (!found) break; // No matching rule found
}
}
if (currentRules.hasOwnProperty("")) { // Fallback rule
fallbackRule = currentRules[""];
}
if (this.#validRule(fallbackRule)) {
return fallbackRule;
}
return null;
}
/**
* Increments the statistics.
* @param {Object} increment The incremental statistics.
* @returns {void}
*/
#incrementStatistics(increment) {
for (const [key, value] of Object.entries(increment)) {
this.#statistics[key] += value;
}
if (typeof CustomEvent === "function") {
this.dispatchEvent(new CustomEvent("statisticschange", {
detail: increment
}));
} else {
this.dispatchEvent(new Event("statisticschange"));
}
}
/**
* Applies the given rule to the given URL object.
* @param {URL} urlObj The URL object to apply the rule to.
* @param {Object} rule The rule to apply.
* @param {Function} logFunc The logger function.
* @returns {Promise<[URL, boolean, Object]>} The new URL object, whether to continue and the mode-specific incremental statistics.
*/
async #applyRule(urlObj, rule, logFunc) {
const mode = rule.mode;
const increment = { ...Purlfy.#zeroStatistics }; // Incremental statistics
const lengthBefore = urlObj.href.length;
const paramsCntBefore = urlObj.searchParams.size;
let shallContinue = false;
switch (mode) { // Purifies `urlObj` based on the rule
case "white": { // Whitelist mode
const newParams = new URLSearchParams();
for (const param of rule.params) {
if (urlObj.searchParams.has(param)) {
newParams.set(param, urlObj.searchParams.get(param));
}
}
urlObj.search = newParams.toString();
break;
}
case "black": { // Blacklist mode
if (!rule.std && !Purlfy.#isStandard(urlObj)) {
logFunc("Non-standard URL search string:", urlObj.search);
break;
}
for (const param of rule.params) {
urlObj.searchParams.delete(param);
}
urlObj.search = urlObj.searchParams.toString();
break;
}
case "param": { // Specific param mode
// Process given parameter to be used as a new URL
let paramValue = null;
for (const param of rule.params) { // Find the first available parameter value
if (urlObj.searchParams.has(param)) {
paramValue = urlObj.searchParams.get(param);
break;
}
}
if (!paramValue) {
logFunc("Parameter(s) not found:", rule.params.join(", "));
break;
}
const dest = Purlfy.#applyActs(paramValue, rule.acts ?? ["url"], logFunc);
if (dest && URL.canParse(dest, urlObj.href)) { // Valid URL
urlObj = new URL(dest, urlObj.href);
} else { // Invalid URL
logFunc("Invalid URL:", dest);
break;
}
shallContinue = rule.continue ?? true;
increment.decoded++;
break;
}
case "regex": { // Regex mode
let newUrl = urlObj.href;
for (let i = 0; i < rule.regex.length; i++) {
const regex = new RegExp(rule.regex[i], "g");
const replace = rule.replace[i];
newUrl = newUrl.replaceAll(regex, replace);
}
newUrl = Purlfy.#applyActs(newUrl, rule.acts ?? [], logFunc);
if (newUrl && URL.canParse(newUrl, urlObj.href)) { // Valid URL
urlObj = new URL(newUrl, urlObj.href);
} else { // Invalid URL
logFunc("Invalid URL:", newUrl);
break;
}
shallContinue = rule.continue ?? true;
break;
}
case "redirect": { // Redirect mode
if (!this.fetchEnabled) {
logFunc("Redirect mode is disabled.");
break;
}
const options = {
method: "HEAD",
redirect: "manual",
headers: rule.headers ?? {}
};
if (rule.ua) {
options.headers["User-Agent"] = rule.ua;
}
let dest = null;
try {
const r = await this.#fetch(urlObj.href, options);
if (r.status >= 300 && r.status < 400 && r.headers.has("location")) {
dest = r.headers.get("location");
} else if (r.url !== urlObj.href) {
dest = r.url; // In case `redirect: manual` doesn't work
}
} catch (e) {
logFunc("Error following redirect:", e);
break;
}
if (dest && URL.canParse(dest, urlObj.href)) {
const prevUrl = urlObj.href;
urlObj = new URL(dest, urlObj.href);
if (urlObj.href === prevUrl) { // No redirection
logFunc("No redirection made.");
break;
}
shallContinue = rule.continue ?? true;
increment.redirected++;
} else {
logFunc("Invalid redirect destination:", dest);
}
break;
}
case "visit": { // Visit mode
if (!this.fetchEnabled) {
logFunc("Visit mode is disabled.");
break;
}
const options = {
method: "GET",
redirect: "manual",
headers: rule.headers ?? {}
};
if (rule.ua) {
options.headers["User-Agent"] = rule.ua;
}
let r, html = null;
try {
r = await this.#fetch(urlObj.href, options);
html = await r.text();
} catch (e) {
logFunc("Error visiting URL:", e);
break;
}
if (r.status >= 300 && r.status < 400 && r.headers.has("location")) {
logFunc("Visit mode, but got redirected to:", r.url);
urlObj = new URL(r.headers.get("location"), urlObj.href);
} else if (r.url !== urlObj.href) { // In case `redirect: manual` doesn't work
logFunc("Visit mode, but got redirected to:", r.url);
urlObj = new URL(r.url, urlObj.href);
} else {
const dest = Purlfy.#applyActs(html, rule.acts?.length ? rule.acts : [String.raw`regex:https?:\/\/.(?:www\.)?[-a-zA-Z0-9@%._\+~#=]{2,256}\.[a-z]{2,6}\b(?:[-a-zA-Z0-9@:%_\+.~#?!&\/\/=]*)`], logFunc);
if (dest && URL.canParse(dest, urlObj.href)) { // Valid URL
urlObj = new URL(dest, urlObj.href);
} else { // Invalid URL
logFunc("Invalid URL:", dest);
break;
}
}
shallContinue = rule.continue ?? true;
increment.visited++;
break;
}
case "lambda": {
if (!this.lambdaEnabled) {
logFunc("Lambda mode is disabled.");
break;
}
try {
const lambda = typeof rule.lambda === "string" ? new Purlfy.#AsyncFunction("url", rule.lambda) : rule.lambda;
rule.lambda = lambda; // "Cache" the compiled lambda function
urlObj = await lambda(urlObj);
shallContinue = rule.continue ?? true;
} catch (e) {
logFunc("Error executing lambda:", e);
}
break;
}
default: {
logFunc("Invalid mode:", mode);
break;
}
}
const paramsCntAfter = urlObj.searchParams.size;
increment.param += (["white", "black"].includes(mode)) ? (paramsCntBefore - paramsCntAfter) : 0;
increment.char += Math.max(lengthBefore - urlObj.href.length, 0); // Prevent negative char count
return [urlObj, shallContinue, increment];
}
/**
* Purifies the given URL based on the rules.
* @param {string} originalUrl The original URL to purify.
* @returns {Promise<Object>} The purified URL and the rule applied.
*/
async purify(originalUrl) {
let increment = { ...Purlfy.#zeroStatistics }; // Incremental statistics of a single purification
let shallContinue = true;
let firstRule = null;
let iteration = 0;
let urlObj;
this.#log("Purifying URL:", originalUrl);
const optionalLocation = typeof location !== 'undefined' ? location.href : undefined;
if (originalUrl && URL.canParse(originalUrl, optionalLocation)) {
urlObj = new URL(originalUrl, optionalLocation);
} else {
this.#log(`Cannot parse URL ${originalUrl}`);
return {
url: originalUrl,
rule: "N/A"
}
}
while (shallContinue && iteration++ < this.maxIterations) {
const logi = (...args) => this.#log(`[#${iteration}]`, ...args);
const protocol = urlObj.protocol;
if (protocol !== "http:" && protocol !== "https:") { // Not a valid HTTP URL
logi(`Not a HTTP URL: ${urlObj.href}`);
break;
}
const hostAndPath = urlObj.host + urlObj.pathname;
const parts = hostAndPath.split("/").filter(part => part !== "");
const rule = this.#matchRule(parts);
if (!rule) { // No matching rule found
logi(`No matching rule found for ${urlObj.href}.`);
break;
}
firstRule ??= rule;
logi(`Matching rule: ${rule.description} by ${rule.author}`);
let singleIncrement; // Incremental statistics for the current iteration
[urlObj, shallContinue, singleIncrement] = await this.#applyRule(urlObj, rule, logi);
for (const [key, value] of Object.entries(singleIncrement)) {
increment[key] += value;
}
logi("Purified URL:", urlObj.href);
}
if (firstRule && originalUrl !== urlObj.href) { // Increment statistics only if a rule was applied and URL has been changed
increment.url++;
this.#incrementStatistics(increment);
}
return {
url: urlObj.href,
rule: firstRule ? `${firstRule.description} by ${firstRule.author}` : "N/A"
};
}
}
if (typeof module !== "undefined" && module.exports) {
module.exports = Purlfy; // Export for Node.js
} else {
this.Purlfy = Purlfy; // Export for browser
}