Performer Auto Rater

:placard: Summary Automatically rates performers in Stash based on how you’ve rated their scenes and images, with optional boosts for engagement signals and favourites.
:link: Repository https://codeberg.org/surging9143/performerAutoRate
:information_source: Source URL https://surging9143.codeberg.page/index.yml
:open_book: Install How to install a plugin?

How it works

Ratings use a Bayesian average: new performers start near your library’s mean and converge toward their true average as you rate more of their content. This prevents a single Top Rated scene from inflating a performer you’ve barely seen.

Optional boosts add points on top of that average. Each boost has a multiplier you set (default 0 = off). Most use a log scale — the bonus grows with engagement but tapers off, so the difference between 1 and 10 plays matters more than between 100 and 110. The favorite boost is a flat addition. The tag boost accumulates once per favorited tag across all scenes and images, so performers whose content matches several of your favorite tags benefit more than those who only match one.

Configure multipliers under Settings → Plugins → Performer Auto Rater. Start small (1–2) and adjust to taste.

Usage

Automatic — runs whenever you update a scene, image, or performer.

Manual — Settings → Plugins → Performer Auto Rater → Rate All Performers. Use this after changing o-counts, which don’t trigger the hook.

Requirements

  • Python 3

  • stashapi (included in the official Stash Docker image)

4 Likes

v2.0.0

  • O-Count Boost — o-counter is now a configurable additive log bonus, instead of a confidence booster (set in plugin settings); defaults to 0
  • Play-Count Boost — new configurable bonus based on scene rewatch count (log scale, same as o-count boost)
  • Favorite Boost — flat point bonus for performers marked as a favourite
  • Performer hook — plugin now triggers on performer updates, so favouriting someone recalculates their rating immediately
1 Like

Favourited Tag Boost (t)

Rewards performers whose scenes and images carry tags you’ve favourited in Stash.

1 Like

Probably other ways to improve this system, my issue right now is that i generally only store videos i like, so they all get pulled to around 7-9, which is a bit boring :stuck_out_tongue: You could tweak things to make it a more even distribution if you’d like.

With the new additions of boosts to the median, you can better make your top performers stand out easily. bringing them up from the “average”

My personal setting are ocount=3, played=1, tag=0 and favorite=10, but this will really vary based on preference, and how large your library is, how good you are at tracking o-counts (how how long of a history you have), how many scenes you have rated etc.

tag boost is a bit undercooked perhaps. but generally, favorite tags you like, then those performers will get rewarded for doing scenes with those tags.

1 Like

I can only see StashGifs in the https://surging9143.codeberg.page/index.yml

Looks like it was removed from the index.

My bad will fix. Release pipeline for stashgifs didnt play nice..

Update: fixed

Could you add play duration as a factor maybe ?

Done!

Example contribution with WATCH_BOOST=1:

Watch Hours Bonus
1h +0.7
10h +2.4
50h +3.9
200h +5.3

At 2, it would be double.

2 Likes

Surely the amount scene of video of one performer you have also +0.1 maybe.

Well, quantity is not necessarily a quality in and of itself.

But if you rate their scenes highly (8-9) and use the other boosts it’s already not that hard to get to 100/100. So i don’t think more boosts are strictly necessary

Update: Merged plugins into a new plugin, with support for galleries and groups aswell.

You can find it on the same index, repurposing this thread for this new version

Deprecated in favor of Stash Auto Rater.