Skip to content

SGN-TS Documentation

SGN-TS extends the SGN streaming pipeline framework with time-series capabilities: offset-based timing, uniformly sampled buffers, multi-channel alignment, and signal processing elements.

Installation

pip install sgn-ts
Optional: PyTorch Support 
pip install sgn-ts[torch]

When PyTorch is installed, the TorchBackend, Converter, and Resampler can use GPU-accelerated operations. Without it, all operations fall back to NumPy.

Developer Installation 
git clone https://git.ligo.org/greg/sgn-ts.git
cd sgn-ts
pip install -e ".[dev]"

Run make to verify your development environment.

Where to Start

  • Tutorial — New to SGN-TS? Build your first time-series pipeline step by step.
  • User Guide — Offsets, buffers, frames, slicing, streaming, plotting.
  • Developer Guide — Writing custom sources, transforms, and sinks.
  • Reference — Built-in sources, transforms, sinks, and auto-generated API documentation.
  • Background — Understand the design: why offsets, power-of-2 rates, and frame alignment.
  • sgn: Base streaming pipeline framework
  • sgn-ligo: LIGO-specific utilities for SGN