Checking for missing content type metadata ...

Checking for non-preferred file/folder path names (may take a long time depending on the number of files/folders) ...

CoDesign: P&ID Knowledge Graph and Decision Traceability Dataset
Authors:
Owners: This resource does not have an owner who is an active HydroShare user. Contact CUAHSI (help@cuahsi.org) for information on this resource.
Type: Resource
Storage: The size of this resource is 206.9 KB
Created: Jul 31, 2026 at 5:21 p.m. (UTC)
Last updated: Jul 31, 2026 at 5:32 p.m. (UTC)
Citation: See how to cite this resource
Sharing Status: Public
Views: 121
Downloads: 21
+1 Votes: Be the first one to 
 this.
Comments: No comments (yet)

Abstract

This dataset contains derived P&ID knowledge graph artifacts prepared from the public DEXPI Public Example PIDs / TrainingTestCases repository. It includes normalized DEXPI/Proteus XML inputs, pyDEXPI JSON model exports, GEXF graph exports where available, a SQLite database representation, schema documentation, sample design decision logs, and sample decision-generated candidate graph exports. The dataset supports research on P&ID knowledge representation, GraphRAG, human-AI design communication, and traceable design-decision logging. The candidate graph files are research artifacts and are not final approved engineering P&ID drawings.

Subject Keywords

Content

README.md

CoDesign Data Package

This folder is organized as the data package for the CUAHSI final report.

Source P&ID Information

  • raw/: source folders copied from the DEXPI public example P&ID training cases.
  • pids/: normalized DEXPI/Proteus XML files used as conversion inputs.
  • sample/: small default PFAS treatment example used by the web app.

Original upstream source: DEXPI Public Example PIDs / TrainingTestCases

  • Repository: https://gitlab.com/dexpi/TrainingTestCases
  • DEXPI 1.3 example PIDs: https://gitlab.com/dexpi/TrainingTestCases/-/tree/master/dexpi%201.3/example%20pids?ref_type=heads

The files in this project are derived or normalized research artifacts created from those public examples; the original P&ID example authorship remains with the upstream DEXPI repository.

Transformed Knowledge Graph Data

  • json/: pyDEXPI JSON model exports.
  • gexf/: NetworkX graph exports for knowledge graph analysis and visualization.
  • schemas/knowledge_graph_schema.json: lightweight schema for the graph/data fields used by this prototype.

Design Decision Dataset

  • design_decisions/: exported CSV/JSON decision logs from human or AI-agent runs.
  • Each row should represent one design question/answer exchange, with responder type stored separately from response content.

Detailed Candidate P&ID Dataset

  • candidates/: generated reviewable candidate P&ID JSON files exported from the web app.
  • schemas/candidate_pid_schema.json: lightweight schema for exported candidate files.

Included Sample Outputs

The repository includes one human-run and one AI-agent-run sample output set for report review:

  • design_decisions/sample_human_decision_log.csv
  • design_decisions/sample_agent_decision_log.csv
  • candidates/sample_human_candidate_pid.json
  • candidates/sample_agent_candidate_pid.json

The human sample demonstrates direct user choices. The agent sample demonstrates AI-default assumptions marked for human review.

Related Resources

The content of this resource can be executed by Park, S. (2026). CoDesign software prototype and web application. GitHub. https://github.com/separk-1/codesign
The content of this resource is derived from DEXPI Initiative. Public Example PIDs / TrainingTestCases. GitLab. https://gitlab.com/dexpi/TrainingTestCases

How to Cite

Park, S. (2026). CoDesign: P&ID Knowledge Graph and Decision Traceability Dataset, HydroShare, http://www.hydroshare.org/resource/8bd5aaec9d954b9bb83fd380172bc3b3

This resource is shared under the Creative Commons Attribution CC BY.

 http://creativecommons.org/licenses/by/4.0/
CC-BY

Comments

There are currently no comments

New Comment

required