The Open Source DAS Ecosystem

Derrick Chambers

Ineris — Institut national de l’environnement industriel et des risques

Galileo Fiber Optic Conference, 2026

What open DAS tools exist?

The Tools: Methodology

The Tools: The Results

The Tools: Growth Over Time

The Tools: Growth Over Time

The Tools: Language and Forge

Python: 76%
GitHub: 96%

The Tools: Dependencies

The Tools: Dependencies

Which projects to build on?

Infrastructure: Licences

  Non-Commercial GPL LGPL MIT Licence permissiveness

Infrastructure: DAS4Whales

DAS4Whales

  • Whale-call detection and spectrograms
  • Call localisation and track building
  • End-to-end marine workflows

github.com/DAS4Whales/DAS4Whales

Infrastructure: Algorithms

Lightguide

  • Adaptive frequency–wavenumber denoising
  • Rust-backed parallel processing
  • Gateway into the Pyrocko ecosystem

github.com/pyrocko/lightguide

DASPy

  • Rich seismology algorithms
  • Curvelet denoising and strain conversion
  • Simple Section and Collection workflow

github.com/HMZ-03/DASPy

Infrastructure: Data Foundations

DASCore

  • Comprehensive dataset management
  • Extensible plugin architecture
  • 20+ versioned format handlers

github.com/DASDAE/dascore

Xdas

  • Xarray-inspired data structures
  • Virtual arrays across thousands of files
  • Stateful and streaming pipelines

github.com/xdas-dev/xdas

What are the gaps and opportunities?

Gaps and Opportunities

Interoperability

Five packages have five I/O layers and five ways to implement similar workflows.

github.com/DASDAE/unidas

Tensor Acceleration

NumPy is shared, but acceleration is bolted on unevenly.

github.com/DASDAE/dasjax

Gaps and Opportunities

Metadata Management

Metadata handling and tracking are uneven across the ecosystem.

github.com/dasdae/dascore

Archive Visualisation

Meaningful multi-resolution views of large archives are still missing.

?

But what about AI?

Gaps: Agent Support

SCIENTIST
>>>
Domain OSS
Algorithms
ML Models

Gaps: Agent Support

SCIENTIST
>>>
Domain OSS
Algorithms
ML Models
AGENT

Gaps: Agent Support

SCIENTIST
>>>
Domain OSS
Algorithms
ML Models
AGENT

Gaps: Agent Support

github.com/dasdae/derzug

Conclusions

  • A young field, many new projects each year
  • Python and GitHub dominate
  • Little connectivity, 5 infrastructure tools emerging
  • Gaps and opportunities:
    • Interoperability, Tensor Acceleration, Visualisation, Agent Support
  • Interoperate rather than re-implement

Thank you


https://github.com/dasdae/


derrick.chambers@ineris.fr


derrick.ja.chambers@gmail.com


github.com/d-chambers

Extra

Xdas vs DASCore: Abstractions

Infrastructure: Requirements

  • I/O for multiple files
  • Processing routines
  • Visualisations
  • More than 1 contributor
  • At least 1 year old and active
  • In a code registry

The Tools: Engineering Signals

The Tools: Dependencies

République Française Ineris