dfd

SYS/download.py

@@ -584,10 +584,15 @@ def _download_direct_file(
         filename = filename.split("?")[0]
     
     # Try to get real filename from Content-Disposition header (HEAD request)
+    content_type = ""
     try:
         with HTTPClient(timeout=10.0) as client:
             response = client._request("HEAD", url, follow_redirects=True)
             content_disposition = response.headers.get("content-disposition", "")
+            try:
+                content_type = str(response.headers.get("content-type", "") or "").strip().lower()
+            except Exception:
+                content_type = ""
             if content_disposition:
                 # Extract filename from Content-Disposition header
                 # Format: attachment; filename="filename.pdf" or filename=filename.pdf
@@ -620,9 +625,36 @@ def _download_direct_file(
             else:
                 filename = suggested
 
-    # Final fallback if we still don't have a good filename
-    if not filename or "." not in filename:
-        filename = "downloaded_file.bin"
+    # If we still don't have an extension, try to infer one from Content-Type.
+    # Never fall back to a generic `.bin` extension.
+    try:
+        has_ext = bool(filename and Path(str(filename)).suffix)
+    except Exception:
+        has_ext = False
+
+    if filename and (not has_ext):
+        ct = (content_type or "").split(";")[0].strip().lower()
+        ext_by_ct = {
+            "application/pdf": ".pdf",
+            "application/epub+zip": ".epub",
+            "application/x-mobipocket-ebook": ".mobi",
+            "image/jpeg": ".jpg",
+            "image/png": ".png",
+            "image/webp": ".webp",
+            "image/gif": ".gif",
+            "text/plain": ".txt",
+            "application/zip": ".zip",
+        }
+
+        if ct in ext_by_ct:
+            filename = f"{filename}{ext_by_ct[ct]}"
+        elif ct.startswith("text/html"):
+            # Guardrail: HTML landing pages should not be downloaded as opaque files.
+            raise DownloadError("URL appears to be an HTML page, not a direct file")
+
+    # Final guardrail: if filename is empty, refuse rather than inventing `download.bin`.
+    if not filename or not str(filename).strip():
+        raise DownloadError("Could not determine filename for URL (no Content-Disposition and no path filename)")
 
     file_path = _unique_path(output_dir / filename)
     progress_bar = ProgressBar()
@@ -684,9 +716,15 @@ def _download_direct_file(
         # For direct file downloads, create minimal info dict without filename as title
         # This prevents creating duplicate title: tags when filename gets auto-generated
         # We'll add title back later only if we couldn't extract meaningful tags
+        ext = ""
+        try:
+            ext = Path(str(filename)).suffix.lstrip(".")
+        except Exception:
+            ext = ""
+
         info = {
-            "id": filename.rsplit(".", 1)[0],
-            "ext": filename.rsplit(".", 1)[1] if "." in filename else "bin",
+            "id": str(filename).rsplit(".", 1)[0] if "." in str(filename) else str(filename),
+            "ext": ext,
             "webpage_url": url,
         }