Improve format selection for audio/video downloads

Refactored format selection logic to explicitly handle audio-only and progressive formats using new flags. Updated GUI to pass format metadata, ensuring correct yt-dlp command construction and more reliable downloads for different format types.

src/core/ytsage_downloader.py

@@ -40,6 +40,8 @@ class DownloadThread(QThread):
         url,
         path,
         format_id,
+        is_audio_only=False,
+        format_has_audio=False,
         subtitle_langs=None,
         is_playlist=False,
         merge_subs=False,
@@ -61,6 +63,8 @@ class DownloadThread(QThread):
         self.url = url
         self.path = Path(path)
         self.format_id = format_id
+        self.is_audio_only = is_audio_only
+        self.format_has_audio = format_has_audio
         self.subtitle_langs = subtitle_langs if subtitle_langs else []
         self.is_playlist = is_playlist
         self.merge_subs = merge_subs
@@ -167,26 +171,19 @@ class DownloadThread(QThread):
 
         # Format selection strategy - use format ID if provided or fallback to resolution
         if self.format_id:
-            # Strip the -drc suffix if present to fix issues with certain audio formats
             clean_format_id = self.format_id.split("-drc")[0] if "-drc" in self.format_id else self.format_id
 
-            # Use a simple heuristic: if format_id contains 'audio' or ends with 'a', treat as audio-only
-            # This avoids the need for Python API metadata extraction
-            is_audio_format = "audio" in clean_format_id.lower() or clean_format_id.endswith("a")
-            
-            # For audio-only formats, don't try to merge with video
-            if is_audio_format:
+            # If the selected format is audio-only, pass it directly.
+            if self.is_audio_only:
                 cmd.extend(["-f", clean_format_id])
                 logger.debug(f"Using audio-only format selection: {clean_format_id}")
+            # If the selected format already includes an audio track (progressive), no merge needed.
+            elif self.format_has_audio:
+                cmd.extend(["-f", clean_format_id])
+                logger.debug(f"Using progressive format with bundled audio: {clean_format_id}")
             else:
                 cmd.extend(["-f", f"{clean_format_id}+bestaudio/best"])
-                logger.debug(f"Using video format selection with audio: {clean_format_id}+bestaudio/best")
-
-            # Determine output format based on the selected format ID - only for video formats
-            # Note: This is a best-effort approach without Python API metadata
-            if not is_audio_format:
-                # Let yt-dlp handle format detection automatically via CLI
-                logger.debug("Letting yt-dlp CLI determine output format automatically")
+                logger.debug(f"Using video-only format merged with best audio: {clean_format_id}+bestaudio/best")
         else:
             # If no specific format ID, use resolution-based sorting (-S)
             res_value = self.resolution if self.resolution else "720"  # Default to 720p if no resolution specified

src/gui/ytsage_gui_format_table.py

@@ -262,6 +262,8 @@ class FormatTableMixin:
             # Column 0: Select Checkbox (Always shown)
             checkbox = QCheckBox()
             checkbox.format_id = str(f.get("format_id", ""))  # type: ignore[reportAttributeAccessIssue]
+            checkbox.is_audio_only = bool((f.get("vcodec") or "none").lower() == "none")  # type: ignore[attr-defined]
+            checkbox.has_audio = bool(f.get("acodec") and f.get("acodec") != "none")  # type: ignore[attr-defined]
             checkbox.clicked.connect(lambda checked, cb=checkbox: self.handle_checkbox_click(cb))
             self.format_checkboxes.append(checkbox)
             checkbox_widget = QWidget()
@@ -427,7 +429,11 @@ class FormatTableMixin:
 
         for checkbox in self.format_checkboxes:
             if checkbox.isChecked():
-                return checkbox.format_id
+                return {
+                    "format_id": checkbox.format_id,
+                    "is_audio_only": getattr(checkbox, "is_audio_only", False),
+                    "has_audio": getattr(checkbox, "has_audio", False),
+                }
         return None
 
     def update_format_table(self, formats) -> None:

src/gui/ytsage_gui_main.py

@@ -802,10 +802,13 @@ class YTSageApp(QMainWindow, FormatTableMixin, VideoInfoMixin):  # Inherit from
             return
 
         # Get selected format
-        format_id = self.get_selected_format()
-        if not format_id:
+        selected_format = self.get_selected_format()
+        if not selected_format:
             self.status_label.setText(_('download.please_select_format'))
             return
+        format_id = selected_format["format_id"]
+        is_audio_only = bool(selected_format.get("is_audio_only"))
+        format_has_audio = bool(selected_format.get("has_audio"))
 
         # Show preparation message
         self.status_label.setText(_('download.preparing'))
@@ -859,6 +862,8 @@ class YTSageApp(QMainWindow, FormatTableMixin, VideoInfoMixin):  # Inherit from
             url=url,
             path=path,
             format_id=format_id,
+            is_audio_only=is_audio_only,
+            format_has_audio=format_has_audio,
             subtitle_langs=selected_subs,  # Pass the list of selected subs
             is_playlist=self.is_playlist,  # Use the flag directly
             merge_subs=self.merge_subs_checkbox.isChecked(),