Geotech AI Assessment Agents (using CrewAI)

This project demonstrates a multi-agent system built with CrewAI to perform a preliminary geotechnical site assessment. The crew consists of specialized AI agents (powered by OpenAI's GPT-4o) that collaborate sequentially to:

1. Analyze project requirements.
2. Review summarized site investigation data.
3. Perform a basic bearing capacity estimation.
4. Compile the findings into a summary report.

This serves as a basic example of applying AI agent collaboration to a typical geotechnical engineering workflow.

Features

• Role-Based Agents: Clearly defined roles for requirements analysis, data review, engineering calculation, and reporting.
• Sequential Workflow: Tasks are executed in a logical order using CrewAI's sequential process.
• Configuration Driven: Agents and Tasks are defined in YAML files (config/agents.yaml, config/tasks.yaml) for easy modification.
• OpenAI Integration: Uses GPT-4o for agent intelligence (requires an API key).
• Markdown Reporting: Generates a final summary report in Markdown format.

Project Structure

• geotech_ai_agents/
  • - .env
  • - pyproject.toml
  • - README.md
  • - src/
    • - geo_assessment_crew/
      • - __init__.py
      • - main.py
      • - crew.py
      • - config/
        • - __init__.py
        • - agents.yaml
        • - tasks.yaml

Prerequisites

• Python 3.10+
• Poetry (recommended for dependency management) or pip
• OpenAI API Key

Setup

1. Clone the repository:

   git clone https://github.com/yaredwb/geotech_ai_agents.git
   cd geotech_ai_agents

2. Create a virtual environment (recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install dependencies:

   • Using Poetry:

     pip install poetry # If you don't have it already
     poetry install

   • Using pip:

     pip install -r requirements.txt # (If you generate one)
     # Or manually:
     # pip install crewai crewai-tools langchain-openai python-dotenv

4. Set up Environment Variables: Create a file named .env in the project root directory (geotech_ai_agents/.env). Add your OpenAI API key to this file:

   OPENAI_API_KEY=sk-YOUR_OPENAI_API_KEY_HERE
   # You can optionally specify the model, though it's also set in agents.yaml
   # OPENAI_MODEL_NAME=gpt-4o
   

   Replace sk-YOUR_OPENAI_API_KEY_HERE with your actual key.

Running the Crew

Ensure your virtual environment is activated. The main script supports several ways to run the application:

Default Mode

Uses built-in sample data for project brief and site information:

python src/geo_assessment_crew/main.py

Interactive Mode

Prompts for project brief and site data inputs with the option to use default values:

python src/geo_assessment_crew/main.py --interactive

File Input Mode

Read project brief and/or site data from text files:

python src/geo_assessment_crew/main.py --brief path/to/brief-file --site-data path/to/site_data-file

Combined Modes

You can combine interactive mode with file inputs. For example, to use a file for the project brief but provide site data interactively:

python src/geo_assessment_crew/main.py --interactive --brief path/to/brief-file

Output

The application generates a markdown report file at output/preliminary_geo_report.md. When the execution completes, you'll be asked if you want to display the report content in the terminal.

Customization

• Modify agent behavior by editing config/agents.yaml
• Adjust task definitions in config/tasks.yaml
• For more advanced customization, you can modify the crew structure in crew.py