🧠 Codebase Assistant (RAG Demo with Gemini)

An intelligent AI assistant built in Python that uses Retrieval-Augmented Generation (RAG) to understand and answer natural language questions about any Python codebase.

You can point it at a local repository (for example, a FastAPI project) — it will analyze and "learn" the code, allowing you to ask questions like:

“How is user authentication handled?”
“Show me the Pydantic model for an Article.”

---

🚀 Tech Stack

• Google Gemini → For powerful Large Language Model (LLM) and embedding generation.
• LangChain → Framework to chain RAG components together.
• FAISS → Fast, local, in-memory vector database for similarity search.

---

⚙️ How It Works (RAG Pipeline)

The assistant follows a classic Retrieval-Augmented Generation pipeline:

1. Load → Scans your repository and loads all .py files.
2. Split → Intelligently splits code into smaller, context-aware chunks.
3. Embed → Converts each chunk into a vector using Google’s embedding-001 model.
4. Store → Saves all vectors into a local FAISS vector database.
5. Retrieve → When you ask a question, it finds the top 5 most relevant code chunks.
6. Augment → Adds these retrieved chunks to your original question as context.
7. Generate → Sends the augmented prompt to Gemini, which produces an accurate answer based on the provided context.

---

🧩 Example Workflow

1. You ask:

   “How is database dependency injection handled?”

2. The assistant retrieves the top 5 related code snippets from your repository.

3. It combines your question + retrieved code snippets into a single prompt.

4. Gemini generates a precise, context-aware answer — grounded only in your codebase.

---

Setup and Installation

Prerequisites

Make sure you have:

• Python 3.9+
• Git
• A Google Gemini API key

---

Clone the Repository

git clone https://github.com/supreetbhat/codebase_assistant.git
cd codebase_assistant

⚙️ Setup and Usage

🧱 Set Up a Virtual Environment

It’s strongly recommended to use a virtual environment.

For macOS/Linux:

python3 -m venv venv
source venv/bin/activate

For Windows:

python -m venv venv
.\venv\Scripts\activate

Install Dependencies

Install all required libraries from requirements.txt

pip install -r requirements.txt

Set Up Your API Key

You’ll need a Google Gemini API key. Get it from Google AI Studio. Then create a .env file:

cp .env.example .env

Open .env and add your key:

GOOGLE_API_KEY="YOUR_API_KEY_HERE"

💡 Note: You must enable billing on your Google Cloud project to use the embedding API.

🚀 How to Run

Run the assistant and pass your target codebase path as an argument:

python code_assistant.py /path/to/your/codebase

Example:

python code_assistant.py ../my-fastapi-project

The script will first index the codebase (this may take a minute). Once you see:

✅ Codebase Assistant is ready!

You can start asking natural language questions. Type exit to quit.

💬 Example Questions to Ask

“What does the /api/users/login endpoint do?”

“How do I create a new article?”

“Show me the Pydantic model for a User.”

“How is database dependency injection handled?”

“What fields are required to create a new user?”

“Explain the logic for following another user.”