FirstAidRAG: A Smart Emergency Response Assistant Powered by LlamaIndex

FirstAidRAG is an intelligent first aid assistant that utilizes LlamaIndex and Ollama to process documents, extract relevant first aid information, and provide real-time assistance.

---

Script Overview

Below is a description of the key scripts used in the project:

• extract_text.py: Extracts structured text from a PDF using LlamaParse API. If LlamaParse fails, it falls back to SimpleDirectoryReader. The script processes the PDF, logs results, and prints a preview of the extracted text.

• clean_text.py: Cleans and formats extracted text, fixing OCR errors, hyphenation issues, and section headers while removing irrelevant content and preserving important terms.

• ingest_data.py: Extracts text from a PDF, cleans it, and saves the cleaned content into a Markdown file. The script processes each section of the document, filters relevant content, and handles file saving.

• index_data.py: Creates or loads a FAISS index for vector-based search using Ollama embeddings. This script processes cleaned text from the Markdown file and stores the index for efficient querying.

• evaluate_rag.py: Evaluates the first aid responses of the LLM based on relevancy and faithfulness scores. Results are saved to a CSV file for further analysis.

• query_engine.py: Loads a FAISS index, retrieves relevant documents using AutoMergingRetriever, and generates a response based on the first aid prompt. If no relevant documents are found, it returns a default message.

• chat_engine.py: Creates a chat engine for interactive first aid assistance using a memory buffer. Streams real-time responses to user input and provides an option to exit the conversation.

---

Setup

Clone the repository and navigate into the project folder:

git clone https://github.com/gabbygab1233/firstaid_assistant.git
cd firstaid_assistant

---

Install all code dependencies

Install the necessary Python packages using pip:

pip3 install -r requirements.txt

---

Configure LlamaParse API Key

1. Signup for LlamaCloud here to get an API key for LlamaParse.
2. After signing up, create an API key and paste it in the config.yaml file as follows:

config.yaml

LLAMA_CLOUD_API_KEY: put-your-api-key-here

---

Install Ollama

Download Ollama from this link. Once downloaded, pull the necessary models:

ollama pull llama3.1
ollama pull nomic-embed-text

Running the Application

Run Ollama in the Background

Start the Ollama server in the background to serve the models:

ollama serve

---

Running the Pipeline

Once Ollama is running, execute the pipeline by running the following command:

python3 main.py

---

Running the Application

To launch the application, run the following command:

chainlit run app.py -w

---