scTimeBench Documentation

Welcome to the scTimeBench documentation!

This project provides a comprehensive benchmarking suite for single cell time series methods. You will find detailed information on how to run the benchmark and how to extend it with new methods, metrics and datasets.

Contents

• Installation & Setup
  • Install: pip
  • Install: UV
  • Python Version
  • Repository Layout
• Default Run 🚦
  • Overview
  • Start from a config file
  • Select the method
  • Select compatible metrics
  • Let the metric choose default datasets
  • Dataset overrides & customization
  • Custom preprocessing
  • Understand run modes
  • Validate paths and outputs
  • Checklist
• Leaderboard 🥇
  • CSV Leaderboard Viewer
  • Leaderboard Plots
• Dataset Contribution 🗂️
  • Overview
  • Steps for Contribution
  • Checklist
• Metric Contribution 🎯
  • Overview
  • Metric Base Classes
  • Metric Hierarchy
  • Dataset Matching
  • Metric Groups
  • Required Outputs
  • Validation
  • Checklist
• Method Contribution 🛠️
  • Overview
  • Implementation Layouts
  • Runner Structure
  • Output Contract
  • Setup Script
  • Configuration
  • Testing
  • Checklist
• API Reference 📒
  • Core Runtime
  • Dataset Infrastructure
  • Method Execution
  • Metric Framework
  • Trajectory Inference

TODO * Update documentation for sqlite database management and csv extraction.