TechX-Resources · PaulPextra · Jun 19, 2025 · Jun 19, 2025 · Jun 20, 2025 · Jun 20, 2025
diff --git a/app/api/conversation/[id]/route.ts b/app/api/conversation/[id]/route.ts
@@ -45,19 +45,14 @@ export async function GET(
   { params }: { params: Promise<{ id: string }> }
 ): Promise<NextResponse> {
   try {
+    // console.log(params);
     await ensureInitialized();
     const id = (await params).id;
-
-    // Get conversation record from database
     const record = await getConversationRecord(id);
+    const signedUrl = await s3Client.getSignedReadUrl(record.contentKey);
+    console.log("Error Debuging", id, record, signedUrl);
 
-    // Get conversation content from S3
-    const content = await s3Client.getConversationContent(record.contentKey);
-
-    return NextResponse.json({
-      conversation: record,
-      content: content,
-    });
+    return NextResponse.json({ url: signedUrl });
   } catch (error) {
     console.error('Error retrieving conversation:', error);
 

diff --git a/app/api/conversation/route.ts b/app/api/conversation/route.ts
@@ -72,9 +72,22 @@ export async function POST(req: NextRequest) {
       return NextResponse.json({ error: '`htmlDoc` must be a file field' }, { status: 400 });
     }
 
-    // Parse the conversation from HTML
+    // Parse the conversation from HTML if skipScraping is false, otherwise skip parsing
+    const skipScraping = formData.has('skipScraping');
+    let conversation;
     const html = await file.text();
-    const conversation = await parseHtmlToConversation(html, model);
+    if (!skipScraping) {
+      conversation = await parseHtmlToConversation(html, model);
+    } else {
+      // Remove CSS rule from the HTML string
+      const cleanedHtml = html.replace(/body\s*\{[^}]*\}/gm, '');
+      conversation = {
+        model: model,
+        content: cleanedHtml,
+        scrapedAt: new Date().toISOString(),
+        sourceHtmlBytes: cleanedHtml.length,
+      };
+    }
 
     // Generate a unique ID for the conversation
     const conversationId = randomUUID();
@@ -162,4 +175,4 @@ export async function GET(req: NextRequest) {
     console.error('Error retrieving conversations:', err);
     return NextResponse.json({ error: 'Internal error, see logs' }, { status: 500 });
   }
-}
+}
diff --git a/lib/parsers/chatgpt.ts b/lib/parsers/chatgpt.ts
@@ -1,13 +1,39 @@
 import type { Conversation } from '@/types/conversation';
+import { JSDOM } from 'jsdom';
 
 /**
- * Extracts a ChatGPT share page into a structured Conversation.
+ * Extract ChatGPT conversation blocks from HTML input without modifying styles.
+ * @param html Raw HTML string
  */
 export async function parseChatGPT(html: string): Promise<Conversation> {
+  const dom = new JSDOM(html);
+  const document = dom.window.document;
+
+  // Grab only the core conversation blocks
+  const blocks = Array.from(
+    document.querySelectorAll('div.markdown.prose.dark\\:prose-invert.w-full.break-words')
+  );
+
+  if (blocks.length === 0) {
+    throw new Error('Conversation content not found');
+  }
+
+  // Leave all <link> styles untouched and embed conversation only
+  const htmlContent = `
+    <html>
+      <head>
+        ${document.head.innerHTML}
+      </head>
+      <body class="dark">
+        ${blocks.map((b) => b.outerHTML).join('\n')}
+      </body>
+    </html>
+  `;
+
   return {
     model: 'ChatGPT',
-    content: html,
+    content: htmlContent,
     scrapedAt: new Date().toISOString(),
-    sourceHtmlBytes: html.length,
+    sourceHtmlBytes: Buffer.byteLength(htmlContent),
   };
 }