feat: add youtube playlist support

[AmanoSpica]

Summary

This pull request adds support for handling YouTube playlist URLs by expanding them into individual video URLs before processing. (fixes #9)

Changes

Backend

• Added a new /api/playlist POST endpoint in app.py
• Executes the command yt-dlp --flat-playlist -J <url> to retrieve playlist entries (without format data)
• Returns the extracted video URLs from playlist entries as a JSON array

Frontend

• Updated templates/index.html to detect YouTube playlist URLs
• Calls the /api/playlist endpoint when a playlist URL is detected
• Replaces the playlist URL with a list of individual video URLs for further processing

Result

• Users can input YouTube playlist URLs directly
• Playlists are automatically expanded into individual video URLs
• Enables seamless processing without manual extraction

Difference from #27

This PR improves the following points compared to #27:

• /api/info processing time was too long

  • Due to yt-dlp behavior, fetching an entire playlist at once is slow
  • This PR expands the playlist beforehand, making it behave the same as multiple individual URLs

• Users could not see progress clearly

  • Since processing now behaves like multiple URLs, each video's result is shown as soon as it is ready
  • This improves user experience (UX)

[Ranjesh2002]

Hey, I tested this locally and noticed it only fetches the first video from a playlist URL rather than all videos. I have a working implementation in #27 that fetches the full playlist by removing --no-playlist from both get_info() and run_download() and parsing the multiple JSON lines yt-dlp returns. Feel free to check it out!

[AmanoSpica]

@Ranjesh2002
Ah, good catch — I was only handling pure playlist URLs (e.g. https://www.youtube.com/playlist?list=xxxx) and missed the case where a video URL includes a playlist (e.g. watch?v=xxxx&list=xxxx).

I've updated the implementation to also detect and expand those cases.

Could you try it again when you have a moment?

[AmanoSpica]

One advantage of expanding the playlist upfront (as in my approach) is that it may make future enhancements easier.

For example:

• The total number of videos is known in advance, which could help when implementing a progress bar
• Each video is handled independently, so it would be easier to support cancellation without needing to terminate a running yt-dlp process

Just wanted to share this as a possible benefit from a future extensibility/UX perspective.

[Ranjesh2002]

Just tested with the watch?v=xxx&list=xxx format too, my implementation in #27 handles it correctly and expands all 13 videos without any additional URL detection logic. Since --no-playlist is removed entirely, yt-dlp naturally expands any URL containing a playlist parameter. No special casing needed!