chore: bump yt-dlp to 2026.6.9

.github/workflows/lint_ruff.yml

@@ -2,7 +2,17 @@ name: Lint Python code
 
 on:
   push:
+    branches:
+      - main
+    paths-ignore:
+      - "assets/**"
+      - "README.md"
   pull_request:
+    branches:
+      - main
+    paths-ignore:
+      - "assets/**"
+      - "README.md"
 
 jobs:
   lint:
@@ -15,4 +25,4 @@ jobs:
         run: pip install ruff
 
       - name: Run linter
-        run: ruff check .
+        run: ruff check .

.github/workflows/unit-tests.yml

@@ -5,21 +5,15 @@ on:
   push:
     branches: 
       - main
-    paths:
+    paths-ignore:
-      - "src/**"
+      - "assets/**"
-      - "tests/**"
+      - "README.md"
-      - "pyproject.toml"
-      - "pytest.ini"
-      - "ytpl-sync-entry.py"
   pull_request:
     branches:
       - main
-    paths:
+    paths-ignore:
-      - "src/**"
+      - "assets/**"
-      - "tests/**"
+      - "README.md"
-      - "pyproject.toml"
-      - "pytest.ini"
-      - "ytpl-sync-entry.py"
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.ref }}

.github/workflows/update-yt-dlp.yml

@@ -78,6 +78,9 @@ jobs:
         if: steps.detect.outputs.needs_update == 'true'
         uses: peter-evans/create-pull-request@v8
         with:
+          # Use a non-GITHUB_TOKEN credential so the resulting PR triggers CI workflows.
+          # Configure secrets.PR_WORKFLOW_TOKEN with contents:write and pull-requests:write.
+          token: ${{ secrets.PR_WORKFLOW_TOKEN || github.token }}
           branch: chore/refresh-yt-dlp
           commit-message: "chore: bump yt-dlp to ${{ steps.detect.outputs.latest_yt_dlp }}"
           title: "chore: bump yt-dlp to ${{ steps.detect.outputs.latest_yt_dlp }}"

README.md

@@ -11,7 +11,7 @@ Local-first YouTube playlist synchronization client.
 
 ## What's Included
 
-- GUI (PySide6) playlist manager + sync runner
+- GUI (PySide6 Essentials) playlist manager + sync runner
 - Scanner (yt-dlp extract-only), diff engine, filesystem scan
 - Safe reordering via two-pass rename, recycle deletions
 - Async download queue with simple retry (yt-dlp Python API)
@@ -30,6 +30,7 @@ Download the latest release from this repo's Releases page and pick one:
 - `ytpl-sync-windows-{version}.zip` / `ytpl-sync-linux-{version}.tar.gz` (no ffmpeg bundled)
 
 ## Configure
+
 Application uses a json config that canbe edited from UI or manually
 
 ```json
@@ -73,7 +74,7 @@ Queue / retry:
 - Run `ytpl-sync.exe` (GUI).
 
 ## Tray
- 
+
 - The app supports minimizing to tray on close if the OS provides a system tray; use the tray icon menu to quit.
 - Tray behavior settings (Settings page):
   - `close_to_tray`: close hides to tray (keeps running).

plans/mp3 metadata plan.md

@@ -0,0 +1,161 @@
+# MP3 Metadata Plan
+
+## Subject Area
+
+- Add MP3 tag writing for downloaded YouTube playlist items.
+- Scope is limited to `.mp3` outputs produced by `audio` mode and the MP3 side of `both` mode.
+- Metadata is sourced from YouTube/yt-dlp and embedded after audio extraction.
+
+## Goal
+
+- Write useful MP3 metadata for downloaded playlist items without affecting video-only downloads.
+- Keep the implementation reliable when optional fields are missing.
+- Preserve successful downloads even when metadata embedding partially fails.
+- Provide a per-playlist setting to enable or disable MP3 metadata embedding.
+
+## Required Metadata
+
+- `title` ← video title
+- `artist` ← uploader, fallback to channel
+- `album` ← album name if present
+- `tracknumber` ← playlist index
+- `date` / `year` ← upload date
+- `comment` ← source URL
+- `genre` ← if available
+- `album_art` ← thumbnail
+
+## Configuration Requirement
+
+- Add a per-playlist setting to turn MP3 metadata embedding on or off.
+- Default should be explicitly defined during implementation; recommended default is `enabled` for new configs.
+- The setting should only affect `.mp3` metadata writing and should not change download selection, extraction, or `.mp4` handling.
+
+## Current Constraints
+
+- The current playlist scan keeps only a minimal item shape: title, video id, and playlist index.
+- The scanner uses flat extraction, which is sufficient for diffing but not for full tag data.
+- MP3 extraction currently transcodes audio but does not write ID3 metadata.
+
+## Implementation Strategy
+
+- Keep playlist diffing fast by retaining the current flat scan for remote playlist structure.
+- Fetch full metadata only for items that are actually going to be downloaded or repaired.
+- Write metadata only after MP3 extraction completes successfully.
+- Treat metadata embedding as a post-processing step that can fail softly without discarding the MP3.
+
+## Work Breakdown
+
+### 1. Extend the metadata model
+
+- Add optional fields to `PlaylistItem` for:
+  - uploader
+  - channel
+  - album
+  - upload_date
+  - genre
+  - thumbnail_url
+  - webpage_url
+- Keep `artist` as a derived value instead of storing a separate field.
+
+### 2. Fetch full per-video metadata
+
+- Introduce a metadata fetch step for each item selected for download.
+- Use yt-dlp per-video extraction to retrieve richer fields than the flat playlist entry provides.
+- Prefer canonical values from the video page payload for upload date, uploader/channel, album, genre, thumbnail, and source URL.
+
+### 3. Carry metadata through the download pipeline
+
+- Ensure the enriched `PlaylistItem` reaches the download job and post-processing stage.
+- Keep this propagation in-memory unless restart-safe metadata persistence becomes necessary later.
+- Avoid changing unrelated sync behavior for video-only items.
+- Carry the per-playlist MP3 metadata enabled/disabled setting into the post-processing step.
+
+### 4. Add an MP3 tag writer
+
+- Add `mutagen` as the ID3 writing dependency.
+- Implement a focused tagging component that maps `PlaylistItem` metadata into ID3 frames.
+- Omit fields when the source value is missing instead of writing placeholders.
+
+### 5. Map fields into ID3 tags
+
+- `title` → video title
+- `artist` → uploader, fallback to channel
+- `album` → album if present
+- `tracknumber` → playlist index
+- `date/year` → parsed upload date
+- `comment` → canonical source URL
+- `genre` → genre if present
+
+### 6. Embed album art
+
+- Download the selected thumbnail for the video after the media download succeeds.
+- Attach thumbnail data as embedded cover art when the image type is supported.
+- Fail soft if thumbnail retrieval or embedding fails, and keep the MP3 intact.
+
+### 7. Integrate into modes
+
+- `audio` mode:
+  - download source media
+  - extract MP3
+  - write MP3 tags only when the setting is enabled
+  - delete temporary/source MP4 if configured
+- `both` mode:
+  - download source media
+  - extract MP3
+  - write MP3 tags only when the setting is enabled
+  - keep MP4 unchanged
+- `video` mode:
+  - no MP3 tagging path
+
+### 8. Add configuration surface
+
+- Add the new per-playlist setting to the playlist config model and default config output.
+- Expose the setting in the playlist configuration UI, not as a global app setting.
+- Keep the naming explicit, for example `write_mp3_metadata` or `embed_mp3_metadata`.
+
+## Error Handling Rules
+
+- If download fails, no tagging runs.
+- If extraction fails, no tagging runs.
+- If metadata embedding is disabled, skip the tagging step entirely.
+- If tagging fails, mark the tag step as failed in logs/events but keep the MP3 file.
+- If thumbnail embedding fails, continue with text metadata only.
+- Missing `album` or `genre` is normal and should not be treated as an error.
+
+## Testing Plan
+
+- Unit test metadata mapping from yt-dlp info to the internal metadata model.
+- Unit test ID3 writing against a temporary MP3 fixture.
+- Unit test fallback behavior:
+  - uploader missing, channel present
+  - album missing
+  - genre missing
+  - thumbnail missing
+- Integration test the audio post-processing path with tagging mocked.
+- Integration test the both-mode MP3 path with tagging mocked.
+
+## Documentation Updates
+
+- Document that MP3 tags are written only for `.mp3` outputs.
+- Document the new per-playlist setting that enables or disables MP3 metadata embedding.
+- Document the field fallback rules, especially artist and album behavior.
+- Document that album art comes from the video thumbnail, not playlist artwork.
+- Document that some YouTube items will not expose album or genre information.
+
+## Dependency Decision
+
+- Recommended library: `mutagen`
+- Reason:
+  - direct ID3 support
+  - reliable field-level control
+  - suitable for embedding cover art
+  - avoids depending on ffmpeg metadata flags for all tag logic
+
+## Delivery Order
+
+- First: add config setting and defaults
+- Second: extend metadata model and add full metadata fetch
+- Third: add MP3 tag writer and field mapping
+- Fourth: add thumbnail embedding
+- Fifth: wire tagging into `audio` and `both`
+- Sixth: add tests and docs

pyproject.toml

@@ -12,8 +12,8 @@ license = { file = "LICENSE" }
 keywords = ["youtube", "yt-dlp", "playlist", "sync"]
 requires-python = ">=3.10"
 dependencies = [
-	"yt-dlp>=2026.3.17",
+	"yt-dlp>=2026.6.9",
-    "PySide6",
+    "PySide6_Essentials>=6.11.1",
 ]
 [project.optional-dependencies]
 test = [