There is a newer version of the record available.

Published April 5, 2026 | Version v1.0.0
Software Open

RCP Framework: A Domain-Agnostic Agentic Architecture for Structured Extraction of Engineering Knowledge

Authors/Creators

  • 1. Division of Product Realization, IEI, Linköping University
  • 2. Saab AB

Description

RCP-Framework v1.0.0

Code and data accompanying the paper:

"A Domain-Agnostic Agentic Architecture for Structured Extraction of Engineering Knowledge"

Submitted to Applied Artificial Intelligence, 2026.

Contents

  • Relational Control Plane (RCP) implementation (SQLite-backed)
  • Case I: Hydroscand hydraulic catalogue extraction pipeline and data
  • Case II: Company B aerospace documentation pipeline (code only; source data subject to confidentiality agreement)
  • Annotated query sets (200 queries, 100 per case study)
  • Evaluation scripts (baseline B1, B2, B3 comparisons)

Requirements

  • Python 3.10+
  • Ollama v0.17.7 with Llama3.2:8b and qwen3-embedding:8b
  • See requirements.txt for dependencies

License

CC-BY 4.0

Notes

If you use this software, please cite it using the metadata below.

Files

oscik559/RCP-Framework-v1.0.0.zip

Files (146.6 MB)

Name Size Download all
md5:6a2fa355938f7c3c747d27b193f1ffa7
146.6 MB Preview Download

Additional details

Related works

Is supplement to
Software: https://github.com/oscik559/RCP-Framework/tree/v1.0.0 (URL)

Software

Repository URL
https://github.com/oscik559/RCP-Framework
Programming language
Python
Development Status
Active

References

  • Ikechukwu, O., Tarkian, M., Nambiar, S., Jonsson, M., & Brax, C. (2026). A Domain-Agnostic Agentic Architecture for Structured Extraction of Engineering Knowledge. Applied Artificial Intelligence.