Improve RSS diff generator (#7)

* Improve RSS diff generator * update dist * improve safety * add test entries * fix void * improve void entry processing * simplify diff * change pubDate of fifth entry * clarify pubDate condition * update comment --------- Co-authored-by: femshima <[email protected]>
utelecon · Dec 26, 2024 · b924d69 · b924d69
1 parent e0e3728
commit b924d69
Show file tree

Hide file tree

Showing 4 changed files with 71 additions and 21 deletions.
diff --git a/dist/index.js b/dist/index.js
@@ -19730,10 +19730,10 @@ Support boolean input list: \`true | True | TRUE | false | False | FALSE\``);
       (0, command_1.issueCommand)("error", (0, utils_1.toCommandProperties)(properties), message instanceof Error ? message.toString() : message);
     }
     exports2.error = error;
-    function warning(message, properties = {}) {
+    function warning2(message, properties = {}) {
       (0, command_1.issueCommand)("warning", (0, utils_1.toCommandProperties)(properties), message instanceof Error ? message.toString() : message);
     }
-    exports2.warning = warning;
+    exports2.warning = warning2;
     function notice(message, properties = {}) {
       (0, command_1.issueCommand)("notice", (0, utils_1.toCommandProperties)(properties), message instanceof Error ? message.toString() : message);
     }
@@ -33316,8 +33316,25 @@ async function tweetRssDiff(rssPaths2, twitterTokens2, safetyLimits2) {
   const parser = new import_rss_parser.default();
   const oldRss = await parseRss(parser, rssPaths2.oldRssPath);
   const newRss = await parseRss(parser, rssPaths2.newRssPath);
-  const oldEntries = new Set(oldRss.items.map((item) => item.link));
-  const posts = newRss.items.filter((item) => !oldEntries.has(item.link)).map((entry) => `${entry.title} ${entry.link}`);
+  const posts = [];
+  for (const entry of newRss.items) {
+    if (!entry.link) {
+      core2.warning(`Entry with title "${entry.title}" has no link.`);
+      continue;
+    }
+    const isOldEntry = oldRss.items.some((oldEntry) => {
+      const diffs = [
+        oldEntry.title !== entry.title,
+        oldEntry.link !== entry.link,
+        oldEntry.pubDate !== entry.pubDate
+      ].filter((d) => d).length;
+      return diffs <= 1;
+    });
+    if (isOldEntry) {
+      break;
+    }
+    posts.push(`${entry.title} ${entry.link}`);
+  }
   if (posts.length === 0) {
     core2.info("No new entry found.");
     return;

diff --git a/src/twitter.ts b/src/twitter.ts
@@ -1,6 +1,6 @@
 import fs from "fs";
 import { setTimeout } from "timers/promises";
-import Parser from "rss-parser";
+import Parser, { type Item } from "rss-parser";
 import { TwitterApi, type TwitterApiTokens } from "twitter-api-v2";
 import * as core from "@actions/core";
 
@@ -16,11 +16,32 @@ export async function tweetRssDiff(
   const oldRss = await parseRss(parser, rssPaths.oldRssPath);
   const newRss = await parseRss(parser, rssPaths.newRssPath);
 
-  const oldEntries = new Set(oldRss.items.map((item) => item.link));
+  const posts = [];
+  for (const entry of newRss.items) {
+    if (!entry.link) {
+      core.warning(`Entry with title "${entry.title}" has no link.`);
+      continue;
+    }
+
+    const isOldEntry = oldRss.items.some((oldEntry) => {
+      const diffs = [
+        oldEntry.title !== entry.title,
+        oldEntry.link !== entry.link,
+        oldEntry.pubDate !== entry.pubDate,
+      ].filter((d) => d).length;
+
+      // If there are more than 1 difference, we assume it's a different entry.
+      return diffs <= 1;
+    });
+    if (isOldEntry) {
+      // Since new entries are always at the top of the feed,
+      // we can stop checking if we reach an old entry.
+      break;
+    }
+
+    posts.push(`${entry.title} ${entry.link}`);
+  }
 
-  const posts = newRss.items
-    .filter((item) => !oldEntries.has(item.link))
-    .map((entry) => `${entry.title} ${entry.link}`);
   if (posts.length === 0) {
     core.info("No new entry found.");
     return;

diff --git a/test/new.xml b/test/new.xml
@@ -6,10 +6,16 @@
     <language>en-us</language>
     <atom:link href="https://www.example.com/rss.xml" rel="self" type="application/rss+xml"/>
     <item>
-      <title>First Entry</title>
-      <link>http://www.example.com/first-entry</link>
-      <description>This is the first entry in the test RSS feed</description>
-      <pubDate>Fri, 25 Oct 2024 00:00:00 GMT</pubDate>
+      <title>Fifth Entry</title>
+      <link>http://www.example.com/second-entry</link>
+      <description>New entry that should be tweeted, even though the link is the same as second entry.</description>
+      <pubDate>Wed, 13 Nov 2024 00:00:00 GMT</pubDate>
+    </item>
+    <item>
+      <title>Fourth Entry</title>
+      <link>http://www.example.com/fourth-entry</link>
+      <description>New entry that should be tweeted</description>
+      <pubDate>Tue, 12 Nov 2024 00:00:00 GMT</pubDate>
     </item>
     <item>
       <title>Second Entry</title>
@@ -20,8 +26,14 @@
     <item>
       <title>Third Entry</title>
       <link>http://www.example.com/third-entry</link>
-      <description>This is the third entry in the test RSS feed</description>
-      <pubDate>Tue, 12 Nov 2024 00:00:00 GMT</pubDate>
+      <description>Since this is after the old entry, this should not be tweeted.</description>
+      <pubDate>Tue, 05 Nov 2024 00:00:00 GMT</pubDate>
+    </item>
+    <item>
+      <title>First Entry</title>
+      <link>http://www.example.com/first-entry</link>
+      <description>This is the first entry in the test RSS feed</description>
+      <pubDate>Fri, 25 Oct 2024 00:00:00 GMT</pubDate>
     </item>
   </channel>
 </rss>
diff --git a/test/old.xml b/test/old.xml
@@ -5,17 +5,17 @@
     <link>http://www.example.com</link>
     <language>en-us</language>
     <atom:link href="https://www.example.com/rss.xml" rel="self" type="application/rss+xml"/>
-    <item>
-      <title>First Entry</title>
-      <link>http://www.example.com/first-entry</link>
-      <description>This is the first entry in the test RSS feed</description>
-      <pubDate>Fri, 25 Oct 2024 00:00:00 GMT</pubDate>
-    </item>
     <item>
       <title>Second Entry</title>
       <link>http://www.example.com/second-entry</link>
       <description>This is the second entry in the test RSS feed</description>
       <pubDate>Fri, 01 Nov 2024 00:00:00 GMT</pubDate>
     </item>
+    <item>
+      <title>First Entry</title>
+      <link>http://www.example.com/first-entry</link>
+      <description>This is the first entry in the test RSS feed</description>
+      <pubDate>Fri, 25 Oct 2024 00:00:00 GMT</pubDate>
+    </item>
   </channel>
 </rss>