Diagram Agent

Table of Contents

• Diagram Agent
  • Table of Contents
  • Overview
  • Features
  • Diagrams Users Can Generate
  • User Experience Flow
  • Front End Features
  • Setup and Run Guidance
    • Prerequisites
    • Docker Version
    • Non-Docker Version
  • Usage
    • Generating Diagrams
    • Generating Questionnaires
  • API Endpoints
  • Example Output
  • Dependencies
  • Contributing
  • Future Developments
  • License

Overview

Diagram Agent is a powerful tool that combines the capabilities of workflow diagramming and questionnaire generation using AI. It leverages the Mermaid markdown-like syntax to create various types of diagrams and uses advanced language models to generate relevant questionnaires.

Features

• Natural Language Processing: Diagram Agent uses advanced language models to generate relevant questionnaires.
• Diagram Generation: Create various types of diagrams such as flowcharts, sequence diagrams, Gantt charts, ER diagrams, class diagrams, and pie charts using natural language descriptions.
• Docker Support: Easily deployable using Docker for consistent environments across different systems.
• Modular Design: Extensible to support additional diagram types.

Diagrams Users Can Generate

• Flowcharts
• Sequence Diagrams
• Gantt Charts
• ER Diagrams
• Class Diagrams
• Pie Charts
• Mindmaps

User Experience Flow

1. Welcome Message: Users are greeted with a welcome message and instructions.
2. Interactive Questionnaire: Users answer questions to specify the diagram details.
3. Diagram Generation: The agent generates the requested diagram based on user input.
4. Review and Modify: Users can request modifications if necessary.
5. Final Output: The final Mermaid code is provided, ready for use in any Mermaid renderer.

Front End Features

• Interactive Console: Real-time interaction via console interface.
• WebSocket Integration: Real-time updates and communication between front end and back end.
• Responsive Design: Accessible on various devices.

Setup and Run Guidance

Prerequisites

• Docker
• Docker Compose

Docker Version

1. Clone the repository:

   git clone https://github.com/your-repo/mermaid-agent.git
   cd mermaid-agent

2. Create a .env file:

   cp sample_env.txt .env

3. Open the .env file and fill in the required values:

   ANTHROPIC_API_KEY="your_anthropic_api_key"
   OPENAI_API_KEY="your_openai_api_key"
   FLASK_APP=app.py
   FLASK_ENV=development
   FLASK_DEBUG=1
   LOG_LEVEL=INFO
   MAX_TOKENS=1000
   TEMPERATURE=0.7
   OUTPUT_DIRECTORY="./outputs"
   SECRET_KEY="your_secret_key"

4. Build and run the Docker container:

   docker-compose up --build

Non-Docker Version

1. Clone the repository:

   git clone https://github.com/your-repo/mermaid-agent.git
   cd mermaid-agent

2. Create a virtual environment and activate it:

   python3 -m venv env
   source env/bin/activate

3. Install the required packages:

   pip install -r requirements.txt

4. Create a .env file:

   cp sample_env.txt .env

5. Open the .env file and fill in the required values:

   ANTHROPIC_API_KEY="your_anthropic_api_key"
   OPENAI_API_KEY="your_openai_api_key"
   FLASK_APP=app.py
   FLASK_ENV=development
   FLASK_DEBUG=1
   LOG_LEVEL=INFO
   MAX_TOKENS=1000
   TEMPERATURE=0.7
   OUTPUT_DIRECTORY="./outputs"
   SECRET_KEY="your_secret_key"

6. Run the Flask application:

   flask run

Access the application at http://localhost:5000.

Usage

Generating Diagrams

Send a POST request to /generate_diagram with the following JSON payload:

{
  "description": "Your diagram description here",
  "diagram_type": "flowchart"
}

Generating Questionnaires

Send a POST request to /generate_questionnaire with the following JSON payload:

{
  "topic": "Your questionnaire topic here"
}

API Endpoints

• GET /: Welcome message
• POST /generate_diagram: Generate a diagram based on the provided description and type
• POST /generate_questionnaire: Generate a questionnaire based on the provided topic

Example Output

Using Diagram Agent to create a flowchart for an AI chatbot system:

flowchart TD
    A[User Input] --> B[API Gateway]
    B --> C{Planner Agent}
    C --> |Create Plan| D[Worker Agent]
    D --> E[Tool Orchestrator]
    E --> F[Web Search]
    E --> G[LLM]
    E --> H[Python REPL]
    E --> I[Chroma DB Vector Store]
    E --> J[Additional Tools]
    F & G & H & I & J --> K[Results Aggregator]
    K --> L[Final Resolver]
    L --> M{QA Agent}
    M --> |Approved| N[Output Publisher]
    M --> |Not Approved| C
    N --> O[User Interface]
    
    subgraph Microservices Architecture
    B
    C
    D
    E
    K
    L
    M
    N
    end
    
    subgraph Data Storage
    P[(Persistent Storage)]
    I --- P
    end
    
    subgraph Security
    Q[Authentication]
    R[Encryption]
    S[Rate Limiting]
    end
    
    B --- Q & R & S

Loading

Dependencies

Diagram Agent relies on several key libraries and frameworks:

Install required packages by running:

pip install -r requirements.txt

Contributing

Contributions are welcome! Please submit a Pull Request.

Future Developments

• Anthropic Prompt caching support (massively reduces the token costs)
• More LLM provider support
• Federated inference capabilities
• Ollama support for local inference
• PDF/CSV upload support during the initial interview questionnaire to provide context from existing documents
• C4 diagram generation support

License

This project is licensed under the MIT License. See the LICENSE file for details.