openBIS Chatbot

A conversational chatbot for querying and retrieving data from openBIS instances using a Large Language Model (LLM). This chatbot allows users to interact with openBIS via natural language, making it easier to explore spaces, projects, experiments, and datasets.

---

Features

• Natural language understanding using Llama 3.2.
• Integration with openBIS API for real-time data retrieval.
• Modular design for easy maintenance and scalability.
• Handles complex queries and provides friendly, human-like responses.
• Error handling for ambiguous or incomplete queries.
• Secure connection to openBIS with robust authentication.

---

Getting Started

Prerequisites

• Python 3.8 or later.
• A functional openBIS instance.
• Access to the Llama 3.2 model via Ollama.
• conda or pip installed on your system.

---

Installation

1. Clone the repository:

   git clone https://github.com/your-repo/openbis-chatbot.git
   cd openbis-chatbot

2. Set up a virtual environment (recommended):

   conda create -n chatbot python=3.8
   conda activate chatbot

3. Install dependencies:

   pip install -r requirements.txt

4. Set up configuration:

   • Edit config/settings.py to include your openBIS URL, username, and password.

---

Usage

1. Run the Chatbot:

   python src/openbis_chatbot/main.py

2. Interact with the Chatbot:

   • Example Queries:
     • "List all projects under the space 'MY_SPACE'."
     • "What datasets are available in 'Project A'?"
     • "Give me all experiments under 'Space X'."

3. Exit the Chatbot:

   • Type exit to close the chatbot.

---

File Structure

openbis-chatbot/
├── .vscode/
│   └── settings.json
├── examples/
│   ├── example_input.json
│   └── example_output.json
├── src/
│   └── openbis_chatbot/
│       ├── utils/
│       │   ├── logger.py
│       │   └── validators.py
│       ├── prompts/
│       │   ├── base_prompt.txt
│       │   └── examples.txt
│       ├── __init__.py
│       ├── main.py
│       ├── data_retrieval.py
│       ├── llm_interface.py
│       ├── parser.py
│       └── prompt.py
├── tests/
│   ├── data/
│   │   └── # here you can put the files in examples/
│   ├── __init__.py
│   ├── test.py
│   ├── test_data_retrieval.py
│   ├── test_llm_interface.py
│   ├── test_parser.py
│   └── test_prompt.py
├── .gitignore
├── README.md
└── pyproject.toml

---

How It Works

1. User Input:

   • The user provides a natural language query.
   • Example: "List all projects under 'MY_SPACE'."

2. LLM Interpretation:

   • The LLM interprets the query and provides structured intent and entities.
   • Output: {"intent": "list_projects", "space": "MY_SPACE"}

3. Query Parsing:

   • The query parser validates and extracts the intent and entities.

4. Data Retrieval:

   • The system uses openBIS APIs to fetch relevant data.

5. Response Generation:

   • The chatbot formats the data into a friendly response and displays it to the user.

---

Testing

To ensure the chatbot is functioning correctly, run the tests:

pytest tests/

---

Dependencies

• pybis: Library for interacting with openBIS.
• ollama-python: Python client for the Llama model.
• httpx: HTTP library for making API requests.

Install all dependencies using:

pip install -r requirements.txt

---

Known Issues

• Ambiguous queries may require follow-up questions for clarification.
• Large datasets may take longer to process.

---

Future Improvements

1. Add support for more openBIS operations (e.g., creating projects, deleting datasets).
2. Enhance the LLM prompt for better context understanding.
3. Integrate with communication platforms like Slack or Teams.

---

Contributing

1. Fork the repository.
2. Create a feature branch:

   git checkout -b feature-name

3. Commit your changes and push:

   git commit -m "Add feature name"
   git push origin feature-name

4. Submit a pull request.

---

License

This project is licensed under the MIT License.

---

Contact

For questions or support, reach out to:

• Name: Carlos Madariaga
• Email: carlosmadariagaaramendi@gmail.com
• GitHub: Carlos Madariaga Github