---
title: LISA Demo
emoji: ⚡
colorFrom: yellow
colorTo: red
sdk: gradio
sdk_version: 4.44.1
app_file: app.py
pinned: false
startup_duration_timeout: 2h
---

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

LISA (Lithium Ion Solid-state Assistant) is a question-and-answer (Q&A) research assistant designed for efficient knowledge management with a primary focus on battery science, yet versatile enough to support broader scientific domains. Built on a Retrieval-Augmented Generation (RAG) architecture, LISA uses advanced Large Language Models (LLMs) to provide reliable, detailed answers to research questions.

DEMO: https://huggingface.co/spaces/Kadi-IAM/LISA

### Installation
1. Clone the Repository:
    ```bash
    git clone "link of this repo"
    cd LISA
    ```

2. Install Dependencies:
    ```bash
    pip install -r requirements.txt
    ```

3. Set enviroment variables (Huggingface for downloading model weights, Tavily for web search and Groq for LLMs):
    ```bash
    export HUGGINGFACEHUB_API_TOKEN=your_api_key_here
    export TAVILY_API_KEY=your_api_key_here
    export GROQ_API_KEY=your_api_key_here
    ```

4. Set Up the Knowledge Base
Populate the knowledge base with relevant documents or research papers. Ensure that documents are in a format (pdf or xml) compatible with the RAG pipeline. By default documents should be located at `data/documents`. After running the following comand, some caches files are saved into `data/db`. ATTENTION: pickle is used to save these caches, be careful with potential security risks.
    ```bash
    python preprocess_documents.py
    ```

4. Running LISA
Once setup is complete, run the following command to launch LISA:
    ```bash
    python app.py
    ```

### About
For more information on our work in intelligent research data management systems, please visit [KadiAI](https://kadi.iam.kit.edu/kadi-ai).