denario

Files (7)

• SKILL.md
• examples.md · references/
• installation.md · references/
• llm_configuration.md · references/
• research_pipeline.md · references/
• demo.cpython-313.pyc · scripts/__pycache__/
• demo.py · scripts/

SKILL.md (first lines)

# Denario

## Overview

Denario is a multiagent AI system designed to automate scientific research workflows from initial data analysis through publication-ready manuscripts. Built on AG2 and LangGraph frameworks, it orchestrates multiple specialized agents to handle hypothesis generation, methodology development, computational analysis, and paper writing.

## When to Use This Skill

Use this skill when:
- Analyzing datasets to generate novel research hypotheses
- Developing structured research methodologies
- Executing computational experiments and generating visualizations
- Conducting literature searches for research context
- Writing journal-formatted LaTeX papers from research results
- Automating the complete research pipeline from data to publication

## Installation

Install denario using uv (recommended):

```bash
uv init
uv add "denario[app]"
```

Or using pip:

```bash
uv pip install "denario[app]"
```

For Docker deployment or building from source, see `references/installation.md`.

## LLM API Configuration

Denario requires API keys from supported LLM providers. Supported providers include:
- Google Vertex AI
- OpenAI
- Other LLM services compatible with AG2/LangGraph

Store API keys securely using environment variables or `.env` files. For detailed configuration instructions including Vertex AI setup, see `references/llm_configuration.md`.

## Core Research Workflow

Denario follows a structured four-stage research pipeline:

### 1. Data Description

Define the research context by specifying available data and tools:

```python
from denario import Denario

den = Denario(project_dir="./my_research")
den.set_data_description("""
Available datasets: time-series data on X and Y
Tools: pandas, sklearn, matplotlib
Research domain: [specify domain]
""")
```

### 2. Idea Generation

Generate research hypotheses from the data description:

```python
den.get_idea()
```

This produces a research question or hypothesis based on the described data. Alternatively, provide a custom idea:

```python
den.set_idea("Custom research hypothesis")
```

### 3. Methodology Development

Develop the research methodology:

```python
den.get_method()
```

This creates a structured approach for investigating the hypothesis. Can also accept markdown files with custom methodologies:

```python
den.set_method("path/to/methodology.md")
```

### 4. Results Generation

Execute computational experiments and generate analysis:

```python
den.get_results()
```

This runs the methodology, performs computations, creates visualizations, and produces findings. Can also provide pre-computed results:

```python
den.set_results("path/to/results.md")
```

### 5. Paper Generation

Create a publication-ready LaTeX paper:

```python
from denario import Journal

den.get_paper(journal=Journal.APS)
```

The generated paper includes proper formatting for the specified journal, integrated figures, and complete LaTeX source.

## Available Journals

Denario supports multiple journal formatting styles:
- `Journal.APS` - American Physical Society format

From lamm-mit/scienceclaw, Apache-2.0.

More from lamm-mit/scienceclaw