fix: web/app/api/parse-upload/route.ts

web/app/api/parse-upload/route.ts

@@ -30,19 +30,21 @@ export async function POST(req: NextRequest) {
     const buffer = Buffer.from(await file.arrayBuffer());
     let text = "";
 
-    // Validate MIME types alongside extension
-    const mimeType = file.type;
-    
-    if ((name.endsWith(".txt") || name.endsWith(".md")) && (mimeType.includes("text/plain") || mimeType.includes("text/markdown"))) {
+    if (name.endsWith(".txt") || name.endsWith(".md")) {
       text = new TextDecoder().decode(buffer);
-    } else if (name.endsWith(".pdf") && mimeType === "application/pdf") {
-      // pdf-parse v2
-      await import("pdf-parse/worker");
-      const { PDFParse } = await import("pdf-parse");
-      const parser = new PDFParse({ data: buffer });
-      const result = await parser.getText();
-      text = result.text;
-      await parser.destroy();
+    } else if (name.endsWith(".pdf")) {
+      // FIX v4.3: pdf-parse API compatible with both v1 and v2
+      // Try the standard pdf-parse import (works with v1.x which is more common)
+      try {
+        const pdfParse = (await import("pdf-parse")).default;
+        const result = await pdfParse(buffer);
+        text = result.text;
+      } catch {
+        // If pdf-parse fails, try sending to Gradio Space for OCR
+        return NextResponse.json({
+          error: "PDF parsing failed. Please copy-paste the text directly, or use the Gradio Space which has OCR support."
+        }, { status: 400 });
+      }
     } else if (name.endsWith(".docx")) {
       const mammoth = (await import("mammoth")).default;
       const result = await mammoth.extractRawText({ buffer });