chore: update node version to 22 and enhance content extraction

Soontao · Jan 5, 2025 · c3f98c3 · c3f98c3
1 parent ea4cb96
commit c3f98c3
Show file tree

Hide file tree

Showing 3 changed files with 16 additions and 3 deletions.
diff --git a/.vscode/launch.json b/.vscode/launch.json
@@ -18,7 +18,7 @@
       "name": "start-rsshub",
       "skipFiles": ["<node_internals>/**"],
       "runtimeExecutable": "node",
-      "runtimeVersion": "20",
+      "runtimeVersion": "22",
       "outputCapture": "std",
       "program": "${workspaceFolder}/packages/rss-server/index.js",
       "envFile": "${workspaceFolder}/.env",
@@ -33,7 +33,7 @@
       "name": "start-rsshub (without ssl)",
       "skipFiles": ["<node_internals>/**"],
       "runtimeExecutable": "node",
-      "runtimeVersion": "20",
+      "runtimeVersion": "22",
       "outputCapture": "std",
       "program": "${workspaceFolder}/packages/rss-server/index.js",
       "envFile": "${workspaceFolder}/.env",

diff --git a/packages/rss-libs/utils/common-utils.js b/packages/rss-libs/utils/common-utils.js
@@ -339,7 +339,15 @@ class GenericEndpointBuilder {
    * @returns {GenericEndpointBuilder} - 当前构建器实例，以便链式调用。
    */
   withContentExtractor(extractor) {
-    this.options.contentExtractor = extractor;
+    if (!this.options.contentExtractor) {
+      this.options.contentExtractor = extractor;
+    } else {
+      const e0 = this.options.contentExtractor;
+      this.options.contentExtractor = async (...args) => {
+        return Object.assign({}, await e0(...args), await extractor(...args))
+      }
+    }
+
     return this;
   }
 

diff --git a/packages/rss-server/customize/cmu-pavlo/index.js b/packages/rss-server/customize/cmu-pavlo/index.js
@@ -1,10 +1,15 @@
 const { GenericEndpointBuilder } = require("rss-libs/utils/common-utils");
+const moment = require("moment");
 
 const endpoint = GenericEndpointBuilder.new()
   .withEndpointPath("/cmu/pavlo")
   .withEntryUrl("https://www.cs.cmu.edu/~pavlo/blog/index.html")
   .withDomAEleLinkExtractor("h4 a")
   .withAutoContentExtractor()
+  .withContentExtractor(($) => ({
+    // 'Posted on October 03, 2016'
+    pubDate: moment($(".postdate",).text()?.slice('Posted on '.length), "MMMM DD, YYYY").toISOString(),
+  }))
   .build();
 
 module.exports = endpoint;