search

giuseppe99barchetta / SuggestArr

Effortlessly request recommended movies, TV shows and anime to Jellyseer/Overseer based on your recently watched content on Jellyfin, Plex or Emby—let SuggestArr handle it all automatically, keeping your library fresh with new and exciting content!
MIT License
227 stars 2 forks source link

General Questions #73

Closed sameleff closed 1 month ago

sameleff commented 1 month ago

I have a few questions about functionality:

  1. How exactly does the recommendation engine work? Which user's viewing history is used when generating a list of media to recommend against (i.e. the plex admin, all plex users, Overseerr admin, Overseer local user if configured)? Are user ratings taken into account?
  2. I was only able to connect to my Plex server via http (insecure) connections. In other apps (e.g. Overseerr, I have been able to connect via https). Will https be enabled in the future?
  3. I like the step-by-step GUI, but once configured it would be nice to be able to edit each individual config rather than stepping through the sequence again from the beginning.
giuseppe99barchetta commented 1 month ago

Hi! Thank you for your interest and for the thoughtful questions about SuggestArr’s functionality. I’ll address each point:

  1. Recommendation Engine: SuggestArr generates recommendations by analyzing the viewing history of all users on the media server. For each user, it retrieves recently viewed content to suggest similar media. Currently, user ratings aren’t factored into the recommendation process, but this could be a future enhancement.
  2. HTTPS Support for Plex: At the moment, SuggestArr supports only HTTP connections for Plex. However, HTTPS support is on my roadmap. I plan to investigate further to identify the cause behind the current SSL limitation for Plex connections.
  3. Configuration Management: I’m glad you appreciate the step-by-step setup interface! I'm already working on making the configuration more flexible, allowing users to adjust individual settings without needing to restart the entire setup process.

Thank you again for the feedback, and I hope you're enjoying SuggestArr!

sameleff commented 1 month ago
  1. Is there a way to restrict which users SuggestArr considers within its recommendation scope? If not, this would be a good feature request candidate. Each user may have different tastes, so one recommendation engine isn't a good match (among other reasons). Also, could the current behavior explain the duplicate recommendations from #74 ? I also experienced this and wasn't sure why, but if SuggestArr pulls viewing history for each user individually and makes recommendations, then if two users have similar tastes it may recommend the same item? Obviously if it combines viewing history from all users before attempting to create issues then that would not explain #74.

Thanks for all the replies - and truly, this is an awesome app. This feature (automatic recommendations based on watch history) is a major gap with a self-hosted server. Definitely exciting!

giuseppe99barchetta commented 1 month ago

Thank you for your detailed feedback and kind words! 😊

Currently, SuggestArr doesn’t have a feature to restrict recommendations to specific users, but it’s a great idea and definitely aligns with enhancing user customization. Implementing a way to filter recommendations based on user preferences or specific user histories would indeed help tailor the results better.

To clarify, SuggestArr is designed to check for existing requests before creating a new one, so theoretically, it should avoid duplicates even if multiple users have similar watch histories. However, it seems that in some cases this doesn’t work as expected, which could lead to duplicate recommendations.

I’ll need to investigate why the duplicate check isn’t always catching these cases. If you have any specific details about the conditions or patterns where duplicates appear, feel free to share them.