Merge pull request #44 from TxtDot/new-match-system

feat: micromatch system to match domains

package-lock.json

@@ -20,12 +20,14 @@
         "fastify": "^4.21.0",
         "ip-range-check": "^0.2.0",
         "jsdom": "^22.1.0",
-        "json-schema-to-ts": "^2.9.2"
+        "json-schema-to-ts": "^2.9.2",
+        "micromatch": "^4.0.5"
       },
       "devDependencies": {
         "@types/ejs": "^3.1.2",
         "@types/express": "^4.17.17",
         "@types/jsdom": "^21.1.1",
+        "@types/micromatch": "^4.0.2",
         "@types/node": "^20.4.10",
         "@typescript-eslint/eslint-plugin": "^6.3.0",
         "@typescript-eslint/parser": "^6.3.0",
@@ -355,6 +357,12 @@
         "@types/node": "*"
       }
     },
+    "node_modules/@types/braces": {
+      "version": "3.0.2",
+      "resolved": "https://registry.npmjs.org/@types/braces/-/braces-3.0.2.tgz",
+      "integrity": "sha512-U5tlMYa0U/2eFTmJgKcPWQOEICP173sJDa6OjHbj5Tv+NVaYcrq2xmdWpNXOwWYGwJu+jER/pfTLdoQ31q8PzA==",
+      "dev": true
+    },
     "node_modules/@types/connect": {
       "version": "3.4.35",
       "resolved": "https://registry.npmjs.org/@types/connect/-/connect-3.4.35.tgz",
@@ -415,6 +423,15 @@
       "version": "7.0.12",
       "license": "MIT"
     },
+    "node_modules/@types/micromatch": {
+      "version": "4.0.2",
+      "resolved": "https://registry.npmjs.org/@types/micromatch/-/micromatch-4.0.2.tgz",
+      "integrity": "sha512-oqXqVb0ci19GtH0vOA/U2TmHTcRY9kuZl4mqUxe0QmJAlIW13kzhuK5pi1i9+ngav8FjpSb9FVS/GE00GLX1VA==",
+      "dev": true,
+      "dependencies": {
+        "@types/braces": "*"
+      }
+    },
     "node_modules/@types/mime": {
       "version": "1.3.2",
       "resolved": "https://registry.npmjs.org/@types/mime/-/mime-1.3.2.tgz",
@@ -889,7 +906,6 @@
     },
     "node_modules/braces": {
       "version": "3.0.2",
-      "dev": true,
       "license": "MIT",
       "dependencies": {
         "fill-range": "^7.0.1"
@@ -1660,7 +1676,6 @@
     },
     "node_modules/fill-range": {
       "version": "7.0.1",
-      "dev": true,
       "license": "MIT",
       "dependencies": {
         "to-regex-range": "^5.0.1"
@@ -2038,7 +2053,6 @@
     },
     "node_modules/is-number": {
       "version": "7.0.0",
-      "dev": true,
       "license": "MIT",
       "engines": {
         "node": ">=0.12.0"
@@ -2243,8 +2257,8 @@
     },
     "node_modules/micromatch": {
       "version": "4.0.5",
-      "dev": true,
+      "resolved": "https://registry.npmjs.org/micromatch/-/micromatch-4.0.5.tgz",
-      "license": "MIT",
+      "integrity": "sha512-DMy+ERcEW2q8Z2Po+WNXuw3c5YaUSFjAO5GsJqfEl7UjvtIuFKO6ZrKvcItdy98dwFI2N1tg3zNIdKaQT+aNdA==",
       "dependencies": {
         "braces": "^3.0.2",
         "picomatch": "^2.3.1"
@@ -2492,7 +2506,6 @@
     },
     "node_modules/picomatch": {
       "version": "2.3.1",
-      "dev": true,
       "license": "MIT",
       "engines": {
         "node": ">=8.6"
@@ -3071,7 +3084,6 @@
     },
     "node_modules/to-regex-range": {
       "version": "5.0.1",
-      "dev": true,
       "license": "MIT",
       "dependencies": {
         "is-number": "^7.0.0"

package.json

@@ -16,12 +16,14 @@
     "fastify": "^4.21.0",
     "ip-range-check": "^0.2.0",
     "jsdom": "^22.1.0",
-    "json-schema-to-ts": "^2.9.2"
+    "json-schema-to-ts": "^2.9.2",
+    "micromatch": "^4.0.5"
   },
   "devDependencies": {
     "@types/ejs": "^3.1.2",
     "@types/express": "^4.17.17",
     "@types/jsdom": "^21.1.1",
+    "@types/micromatch": "^4.0.2",
     "@types/node": "^20.4.10",
     "@typescript-eslint/eslint-plugin": "^6.3.0",
     "@typescript-eslint/parser": "^6.3.0",

src/handlers/google.ts

@@ -3,12 +3,18 @@ import { IHandlerOutput } from "./handler.interface";
 import { EngineParseError } from "../errors/main";
 
 export default async function google(
-  window: DOMWindow
+  window: DOMWindow,
 ): Promise<IHandlerOutput> {
   const googleAnchors = [
     ...window.document.querySelectorAll("a[jsname=ACyKwe]"),
   ] as HTMLAnchorElement[];
 
+  if (!googleAnchors) {
+    throw new EngineParseError(
+      "Failed to find anchors in search result [google]",
+    );
+  }
+
   const results = googleAnchors
     .map((a: HTMLAnchorElement): GoogleProps => {
       const parsedHref = new URL(new URL(a.href).searchParams.get("url")!);
@@ -20,12 +26,6 @@ export default async function google(
     })
     .filter((a) => a.heading);
 
-  if (!googleAnchors) {
-    throw new EngineParseError(
-      "Failed to find anchors in search result [google]"
-    );
-  }
-
   const convertToFormat = (result: GoogleProps, isHtml: boolean) => {
     return isHtml
       ? `<p><a href="${result.href}">${result.siteName} - ${result.heading}</p>`
@@ -41,12 +41,12 @@ export default async function google(
   });
 
   const search = window.document.getElementById(
-    "APjFqb"
+    "APjFqb",
   ) as HTMLTextAreaElement;
 
   const navLinks = [
     ...window.document.querySelectorAll(
-      "table[class=AaVjTc] > tbody > tr > td > a"
+      "table[class=AaVjTc] > tbody > tr > td > a",
     ),
   ].map((l) => {
     const link = l as HTMLAnchorElement;
@@ -81,6 +81,15 @@ export default async function google(
   };
 }
 
+export const GoogleDomains = [
+  "google.*",
+  "google.co.*",
+  "google.com.*",
+  "www.google.*",
+  "www.google.co.*",
+  "www.google.com.*",
+];
+
 interface GoogleProps {
   href: string;
   siteName: string;

src/handlers/main.ts

@@ -5,22 +5,21 @@ import { JSDOM } from "jsdom";
 import { DOMWindow } from "jsdom";
 
 import readability from "./readability";
-import google from "./google";
+import google, { GoogleDomains } from "./google";
-import stackoverflow from "./stackoverflow/main";
+import stackoverflow, { StackOverflowDomains } from "./stackoverflow/main";
 
 import { generateProxyUrl } from "../utils/generate";
 import isLocalResource from "../utils/islocal";
 
-import {
+import micromatch from "micromatch";
-  LocalResourceError,
+
-  NotHtmlMimetypeError,
+import { LocalResourceError, NotHtmlMimetypeError } from "../errors/main";
-} from "../errors/main";
 
 export default async function handlePage(
   url: string, // remote URL
   requestUrl: URL, // proxy URL
   engine?: string,
-  redirect_path: string = "get"
+  redirect_path: string = "get",
 ): Promise<IHandlerOutput> {
   const urlObj = new URL(url);
 
@@ -39,7 +38,12 @@ export default async function handlePage(
 
   [...window.document.getElementsByTagName("a")].forEach((link) => {
     try {
-      link.href = generateProxyUrl(requestUrl, link.href, engine, redirect_path);
+      link.href = generateProxyUrl(
+        requestUrl,
+        link.href,
+        engine,
+        redirect_path,
+      );
     } catch (_err) {
       // ignore TypeError: Invalid URL
     }
@@ -49,25 +53,41 @@ export default async function handlePage(
     return engines[engine](window);
   }
 
-  return fallback[urlObj.host]?.(window) || fallback["*"](window);
+  for (let match of fallback) {
+    if (micromatch.isMatch(urlObj.hostname, match.pattern)) {
+      return match.engine(window);
+    }
+  }
+
+  return engines.readability(window);
 }
 
 interface Engines {
   [key: string]: EngineFunction;
 }
 
-type EngineFunction = (window: DOMWindow) => Promise<IHandlerOutput>;
-
 export const engines: Engines = {
   readability,
   google,
   stackoverflow,
 };
 
+type EngineFunction = (window: DOMWindow) => Promise<IHandlerOutput>;
+export type EngineMatch = {
+  pattern: string | string[];
+  engine: EngineFunction;
+};
+export type EnginesMatch = EngineMatch[];
+
 export const engineList: string[] = Object.keys(engines);
 
-const fallback: Engines = {
+export const fallback: EnginesMatch = [
-  "stackoverflow.com": engines.stackoverflow,
+  {
-  "www.google.com": engines.google,
+    pattern: GoogleDomains,
-  "*": engines.readability,
+    engine: engines.google,
-};
+  },
+  {
+    pattern: StackOverflowDomains,
+    engine: engines.stackoverflow,
+  },
+];

src/handlers/stackoverflow/main.ts

@@ -4,7 +4,7 @@ import { EngineParseError } from "../../errors/main";
 import qPostsHandler from "./questions-posts";
 
 export default async function stackoverflow(
-  window: DOMWindow
+  window: DOMWindow,
 ): Promise<IHandlerOutput> {
   const url = new URL(window.location.href);
 
@@ -29,3 +29,14 @@ export default async function stackoverflow(
 
   return result;
 }
+
+export const StackOverflowDomains = [
+  "stackoverflow.com",
+  "*.stackoverflow.com",
+  "*.stackexchange.com",
+  "askubuntu.com",
+  "stackapps.com",
+  "mathoverflow.net",
+  "superuser.com",
+  "serverfault.com",
+];