YAPT Benchmark with Snakemake

This project automates the execution of rendering scenes using YAPT, collects the resulting .exr images, and extracts statistics (such as RGB channel averages) using oiiotool. The workflow is managed by Snakemake and configured via a CSV file that defines each job's parameters.

📁 Project Structure

.
├── Snakefile               # Snakemake workflow
├── config.yaml            # Configuration file with paths
├── params.csv             # Scene rendering parameters
├── maps/                  # Files mapping scene name to .exr filename
├── images/                # Rendered .exr images
├── stats/                 # Extracted statistics per image
└── results/
    └── stats_summary.csv  # Summary of means for all images

---

Installation

1. Clone this repository

git clone https://github.com/your-username/yapt-sm.git
cd yapt-sm

2. Set up a Python environment

python3 -m venv snakemake
source snakemake/bin/activate
pip install --upgrade pip
pip install snakemake pandas matplotlib

3. Install YAPT and OIIO

• YAPT: Clone and build yapt
• OpenImageIO (oiiotool)
  On Ubuntu:

  sudo apt install openimageio-tools

---

Configuration

Example config.yaml file:

yapt_path: /path/to/yapt/bin
function_path: functions
output_image_path: images

Example params.csv file:

This file defines which scenes to run and with which parameters:

name,source,spp
test1,scene1.ypt,10
test2,scene2.ypt,20

---

Running the Workflow

Once configured, run the full pipeline with:

snakemake --cores 4 --config params_file=params.csv

---

📊 Output

The file results/stats_summary.csv will contain the average value (extracted from the image stats) for each scene:

name,mean
test1,0.791
test2,0.812

---

Cleaning Up

To remove all generated files and reset the workflow:

snakemake --delete-all-output --delete-temp-output