From b25a779d98e4ed674496e6184b424e0a594f2364 Mon Sep 17 00:00:00 2001
From: Pavel Korytov <thexcloud@gmail.com>
Date: Thu, 8 Aug 2024 01:27:33 +0300
Subject: [PATCH] [TldrTechBridge] Fix bridge (#4187)

* [TldrTechBridge] Fix bridge

* yup

---------

Co-authored-by: Dag <me@dvikan.no>
---
 bridges/TldrTechBridge.php | 64 ++++++++++++++++++++++----------------
 1 file changed, 38 insertions(+), 26 deletions(-)

diff --git a/bridges/TldrTechBridge.php b/bridges/TldrTechBridge.php
index d29553479ba..6c96dff7dd3 100644
--- a/bridges/TldrTechBridge.php
+++ b/bridges/TldrTechBridge.php
@@ -1,12 +1,12 @@
 <?php
 
+declare(strict_types=1);
+
 class TldrTechBridge extends BridgeAbstract
 {
     const MAINTAINER = 'sqrtminusone';
     const NAME = 'TLDR Tech Newsletter Bridge';
     const URI = 'https://tldr.tech/';
-
-    const CACHE_TIMEOUT = 3600; // 1 hour
     const DESCRIPTION = 'Return newsletter articles from TLDR Tech';
 
     const PARAMETERS = [
@@ -41,41 +41,53 @@ public function collectData()
     {
         $topic = $this->getInput('topic');
         $limit = $this->getInput('limit');
-        $url = self::URI . $topic . '/archives';
-        $html = getSimpleHTMLDOM($url);
-        $entries_root = $html->find('div.content-center.mt-5', 0);
-        $added = 0;
+
+        $url = self::URI . 'api/latest/' . $topic;
+        $response = getContents($url, [], [], true);
+        $location = $response->getHeader('Location');
+        $locationUrl = Url::fromString($location);
+
+        $this->extractItem($locationUrl);
+
+        $archives_url = self::URI . $topic . '/archives';
+        $archives_html = getSimpleHTMLDOM($archives_url);
+        $entries_root = $archives_html->find('div.content-center.mt-5', 0);
         foreach ($entries_root->children() as $child) {
             if ($child->tag != 'a') {
                 continue;
             }
-            // Convert /<topic>/2023-01-01 to unix timestamp
-            $date_items = explode('/', $child->href);
-            $date = strtotime(end($date_items));
-            $item_url = self::URI . ltrim($child->href, '/');
-            try {
-                $this->items[] = [
-                    'uri'       => self::URI . $child->href,
-                    'title'     => $child->plaintext,
-                    'timestamp' => $date,
-                    'content'   => $this->extractContent($item_url),
-                ];
-            } catch (HttpException $e) {
-                continue;
-            }
-            $added++;
-            if ($added >= $limit) {
+            $this->extractItem(Url::fromString(self::URI . $child->href));
+            if (count($this->items) >= $limit) {
                 break;
             }
         }
     }
 
+    private function extractItem(Url $url)
+    {
+        $pathParts = explode('/', $url->getPath());
+        $date = strtotime(end($pathParts));
+        try {
+            [$content, $title] = $this->extractContent($url);
+
+            $this->items[] = [
+                'uri'       => (string) $url,
+                'title'     => $title,
+                'timestamp' => $date,
+                'content'   => $content,
+            ];
+        } catch (HttpException $e) {
+            // archive occasionally returns broken URLs
+            return;
+        }
+    }
+
     private function extractContent($url)
     {
-        $html = getSimpleHTMLDOM($url);
+        $html = getSimpleHTMLDOMCached($url);
         $content = $html->find('div.content-center.mt-5', 0);
         if (!$content) {
-            throw new HttpException('Could not find content', 500);
+            throw new \Exception('Could not find content');
         }
         $subscribe_form = $content->find('div.mt-5 > div > form', 0);
         if ($subscribe_form) {
@@ -112,7 +124,7 @@ private function extractContent($url)
                 }
             }
         }
-
-        return $content->innertext;
+        $title = $content->find('h2', 0);
+        return [$content->innertext, $title->plaintext];
     }
 }