HCI Robot Path Explanation Project

This project demonstrates how to use a Large Language Model (LLM) to explain and reason about robot navigation paths, integrating environmental context and user questions.

Prerequisites

• Python 3.8+ (recommended: 3.11)
• Ollama (local LLM server)
• llama3.2 model (or another compatible model installed in Ollama)
• make (optional, for easier automation)
• git (for cloning the repository)

How to Download and Launch the Project on Any Computer

1. Clone the repository:

   git clone https://github.com/Vlor999/HCI.git
   cd HCI

2. Install Ollama and the LLM model:

   • Download and install Ollama for your OS (Windows, macOS, Linux).

   • Start Ollama (if not already running):

     ollama serve

   • Pull the model (e.g., llama3.2):

     ollama pull llama3.2

3. Initialize the project structure and install Python dependencies:

   make init
   make install

   This will:

   • Create necessary folders (src/, tests/, data/, log/, etc.)
   • Install Python dependencies in a virtual environment

4. (Optional) Format the code:

   make format

5. Run the project:

   make run

   or manually:

   .venv/bin/python main.py

6. Run the tests:

   make test

7. (Optional) Run coverage and view the HTML report:

   make coverage
   # Then open htmlcov/index.html in your browser

Usage

• When running, the program will display the current robot path and context.
• You can ask multiple questions about the path and its conditions.
• Type exit or quit to end the session.
• After exiting, a Markdown log of the conversation will be saved in the log/ directory.

Project Structure

src/           # Source code (robotPathExplanation.py, path.py, io_console.py, etc.)
tests/         # Unit tests
data/          # Example path data (JSON)
log/           # Conversation logs (Markdown)
doc/           # Documentation and roadmap
evaluation/    # Evaluation scripts and results
Makefile       # Automation commands
requirements.txt

Notes

• You can edit data/paths.json to add or modify path scenarios.
• The LLM model name can be changed in the code if you use a different one.
• The .gitignore file ensures that data, logs, and virtual environments are not committed.

Troubleshooting

• Ollama not running: Make sure ollama serve is active and the model is pulled.
• Port conflicts: Only one Ollama server should run at a time on port 11434.
• Python errors: Ensure you are using the virtual environment (.venv).

---

For more details, see the documentation in doc/ or the comments in the source files.

---

License

This project is licensed under the MIT License.