pysegy

pysegy is a minimal Python library for working with SEGY Rev 1 data. The project provides helpers to read and write files as well as utilities to scan large surveys without loading every trace in memory.

Capabilities

• Read complete SEGY files with segy_read and access both binary and trace headers.
• Write new data sets using segy_write from NumPy arrays.
• Lazily inspect large archives via segy_scan and the SegyScan object.
• Retrieve individual header fields with automatic scaling through get_header.
• Compatible with any fsspec filesystem for local or remote storage.

Installation

Install the project in editable mode from the repository root:

python -m pip install -e .

Or to install the latest pypi release

pip install pysegy

Testing

Run the unit tests with pytest:

pytest -vs

The tests run automatically on GitHub Actions with coverage reports uploaded to Codecov.

Read benchmark

The read benchmark compares median throughput, speedup, and parallel efficiency for multiple process counts:

python benchmarks/read_benchmark.py /path/to/large.segy --workers 1 2 4 8

Use a representative large file and repeat the benchmark on the target storage system. Small files commonly run faster with one worker because process startup and inter-process transfer costs outweigh parallel decoding.

Inspiration

This project started as a lightweight port of the Julia package SegyIO.jl. The goal is to provide a similar user experience for Python while keeping the code base small and easy to understand.