From 6dca8b7d556bc2741072d49d5974dc698bd29f94 Mon Sep 17 00:00:00 2001
From: Theo Sun <theo.sun.github@outlook.com>
Date: Wed, 29 May 2024 23:08:44 +0800
Subject: [PATCH] chore: better support for bing

---
 .vscode/settings.json |  3 ++-
 src/index.mjs         | 41 +++++++++++++++++++++++++++++++++++++----
 src/utils.mjs         | 43 +++++++++++++++++++++++++------------------
 3 files changed, 64 insertions(+), 23 deletions(-)

diff --git a/.vscode/settings.json b/.vscode/settings.json
index b90e0f6..389c87a 100644
--- a/.vscode/settings.json
+++ b/.vscode/settings.json
@@ -7,6 +7,7 @@
     "source.organizeImports": "explicit"
   },
   "cSpell.words": [
-    "Sogou"
+    "Sogou",
+    "weixin"
   ]
 }
\ No newline at end of file
diff --git a/src/index.mjs b/src/index.mjs
index bc749b0..b2afd11 100644
--- a/src/index.mjs
+++ b/src/index.mjs
@@ -21,14 +21,47 @@ app.get(
   }),
 );
 
+app.get(
+  "/weixin",
+  createCommonSearchAPI({
+    urlPrefix: "https://weixin.sogou.com/weixin?type=2&query=",
+    resultsItemSelector: ".news-box .news-list li",
+    titleSelector: "h3 a",
+    linkSelector: "h3 a",
+    descriptionSelector: ".txt-info",
+  }),
+);
+
+app.get(
+  "/zhihu",
+  createCommonSearchAPI({
+    urlPrefix: "https://sogou.com/web?insite=zhihu.com&query=",
+    resultsItemSelector: ".results .vrwrap",
+    titleSelector: "h3 a",
+    linkSelector: "h3 a",
+    descriptionSelector: ".str-text-info",
+  }),
+);
+
+app.get(
+  "/baike",
+  createCommonSearchAPI({
+    urlPrefix: "https://sogou.com/web?insite=baike.baidu.com&query=",
+    resultsItemSelector: ".results .vrwrap",
+    titleSelector: "h3 a",
+    linkSelector: "h3 a",
+    descriptionSelector: ".str-text-info",
+  }),
+);
+
 app.get(
   "/bing",
   createCommonSearchAPI({
     urlPrefix: "https://cn.bing.com/search?q=",
-    resultsItemSelector: "#b_results .b_algo",
-    titleSelector: ".b_tpcn a",
-    linkSelector: ".b_tpcn a",
-    descriptionSelector: ".tptxt",
+    resultsItemSelector: "#b_content #b_results li.b_algo",
+    titleSelector: "h2 a",
+    linkSelector: "h2 a",
+    descriptionSelector: ".b_caption, p.b_algoSlug",
   }),
 );
 
diff --git a/src/utils.mjs b/src/utils.mjs
index 2fb6318..ffde0be 100644
--- a/src/utils.mjs
+++ b/src/utils.mjs
@@ -42,17 +42,16 @@ export function createCommonSearchAPI(options) {
 
     await page.goto(`${options.urlPrefix}${search}`, {
       waitUntil: "networkidle0",
-      timeout: 30_000,
-      referer: "https://cn.bing.com/",
+      timeout: 10_000,
+      referer: options.urlPrefix,
     });
 
     const cards = await page.$$(options.resultsItemSelector);
 
     const searchResults = await Promise.all(
-      cards.map(async (card) => {
-        const item = await card.evaluate((node, options) => {
-          const linkEle = node.querySelector(options.linkSelector);
-          const link = linkEle?.href;
+      cards.map((card) => {
+        return card.evaluate((node, options) => {
+          const link = node.querySelector(options.linkSelector)?.href;
           const title = node.querySelector(options.titleSelector)?.innerText;
           const description = node.querySelector(
             options.descriptionSelector,
@@ -63,11 +62,22 @@ export function createCommonSearchAPI(options) {
             description,
           };
         }, options);
-        if (!item?.link) return;
-        if (!process.env.TF_URL) {
-          console.warn("TF_URL is not set, skipping text extraction");
-          return item;
-        }
+      }),
+    );
+
+    const validSearchResults = searchResults
+      .filter((i) => i.link)
+      .slice(0, parseInt(req.query.top ?? 100));
+
+    await browser.close();
+
+    if (!process.env.TF_URL) {
+      console.warn("TF_URL is not set, skipping text extraction");
+      return validSearchResults;
+    }
+
+    await Promise.all(
+      validSearchResults.map(async (item) => {
         try {
           const res = await axios.post(
             process.env.TF_URL + "/extract",
@@ -75,23 +85,20 @@ export function createCommonSearchAPI(options) {
             {
               headers: {
                 "Content-Type": "application/json",
-                "User-Agent": defaultUserAgent(),
               },
-              timeout: 5_000,
+              timeout: 2_000,
             },
           );
           if (res.status < 300 && res.data?.text) {
-            return { ...item, text: res.data };
+            item.text = res.data.text;
           }
         } catch (error) {
           console.error("Failed to extract text", error.message, item.link);
         }
-
-        return item;
       }),
     );
-    await browser.close();
-    return res.json(searchResults.filter(Boolean));
+
+    return res.json(validSearchResults.filter(Boolean));
   });
 }