Rocco — AI Curator for Dataset Descriptions

AI-powered dataset description evaluation & enhancement framework for the Digital Porous Media Portal.

Rubric-Based Evaluation

Score any dataset description against 10 domain-specific criteria.

Evaluator

RAG-Powered Enhancement

Improve descriptions with relevant excerpts drawn from your papers.

Writer

Iterate

Refine with interactive feedback across multiple rounds until your description is publication-ready.

Streamlit App

---

Getting Started Locally

Quick Install

git clone --branch v1.0.0 --depth 1 https://github.com/digital-porous-media/dpm_rocco_curator.git
cd dpm_rocco_curator
pip install .

Configure LLM Endpoints

cp .env.example .env

# Edit .env to set LLM_PROVIDER, LLM_API_KEY, LLM_MODEL.
# Rocco supports OpenAI-compatible APIs (OpenAI, Anthropic, Gemini, HuggingFace, Ollama, and more!)

Run Rocco UI

streamlit run rocco_ui.py

# Opens at http://localhost:8501

For a full walkthrough, see Quick Start.

---

Getting Started

• Installation
• Quick Start

Description Curator

• Streamlit App
• Evaluator
• Document RAG
• Writer

AI Providers

• Configuration
• Next Steps

Developer Guide

• Architecture
• Contributing
• API Reference
• Prompt Reference