StashVideohasherNode

Plugins Other Projects
,
1
:placard: Summary This is a very simple script that can be run on multiple systems to process a large Stash import of scenes. Instead of running cover and phash generation tasks on the Stash server itself, this script will allow you to do the same thing on as many computers as you would like, with all of the nodes contributing back to the Stash server.
:link: Repository https://github.com/Darklyter/StashVideohasherNode

:videocassette: StashVideohasherNode Script

This is a lightweight script designed to distribute the processing of large Stash scene imports across multiple systems. Instead of burdening the Stash server with tasks like cover generation, scrubber sprite creation, preview video rendering, and perceptual hash (phash) computation, this script enables multiple nodes to contribute those tasks back to the server.

:gear: Requirements

:brain: How It Works

  • Processes scenes in batches of 25 per node
  • Tags batches as "In Process" to prevent duplication
  • Continues processing until fewer than 25 scenes remain
  • Scenes are selected based on missing phash
  • If phash is generated but other tasks fail (e.g., cover image), the scene won’t be reprocessed

:speech_balloon: Support

The script is well-commented. For questions, reach out via Discord (if you know this script, you probably know how to find me there).

:hammer_and_wrench: CLI Options

You can run the script with the following options, all of which can also be set in config.py:

usage: phash_videohasher_main.py [-h] [--windows] [--generate-sprite] [--generate-preview]
                                 [--batch-size BATCH_SIZE] [--max-workers MAX_WORKERS]
                                 [--dry-run] [--verbose] [--once]

Stash Scene Processor CLI

options:
  -h, --help              Show this help message and exit
  --windows               Use Windows-style paths and binaries
  --generate-sprite       Enable sprite image generation
  --generate-preview      Enable preview video generation
  --batch-size BATCH_SIZE Number of scenes to process per run (default: 25)
  --max-workers MAX_WORKERS Number of threads for parallel processing (default: 4)
  --dry-run               Simulate processing without writing changes
  --verbose               Enable detailed output and progress bars
  --once                  Run a single batch and exit