Building a Mini RAG Application with FastAPI + ChromaDB (WHO Report Q&A Project)

Описание: Project Overview

This project implements a minimal Retrieval-Augmented Generation (RAG) API using FastAPI that answers questions from the WHO World Health Statistics 2024 report.

The pipeline includes:
1. Parsing the PDF → creating vector indexes → enabling retrieval
2. Querying ChromaDB to fetch relevant context
3. Asynchronously calling Google Gemini to generate final answers
A small load-testing script is also included, which fires 5 concurrent queries against the API and measures total execution time.

What’s Inside
1. Minimal FastAPI server (main.py) exposing a /query endpoint
2. Document flow: PDF parsing → vector indexing → retrieval
3. Async integration with Gemini for generation

Tech Stack
1. FastAPI + Uvicorn (API server)
2. ChromaDB (vector store)
3. LlamaIndex (for indexing and retrieval orchestration)
4. Google Gemini (google-genai) (LLM for final answer generation)

Resources
GitHub Repository: https://github.com/GenAIApplication/r...
Code Architecture Reference: You can explore a detailed explanation of the architecture on Codalogy
under the project Simple RAG. https://codalogy.com

#RAG #FastAPI #LLM #GenAI #GoogleGemini #ChromaDB #LlamaIndex #VectorSearch #APIDevelopment #RetrievalAugmentedGeneration #Python #OpenSource #MachineLearning #Codalogy #WHOData #AsyncPython