Gemi Chat 🤖

A CLI-based AI chatbot for analyzing UBER SEC 10-K filings using Retrieval-Augmented Generation (RAG) and Google's Gemini API.

🚀 Features

• AI-Powered Analysis: Interactive chat with a FunctionAgent that answers questions about UBER's financial data from 2019-2022.
• RAG with LlamaIndex: Indexes SEC filings for accurate, context-aware responses.
• CLI Interface: Simple command-line tools for data loading and chatting.
• Modular Design: Clean, maintainable code structure for easy extension.

📦 Installation

1. Clone the repo:

   git clone https://github.com/yourusername/gemi-chat.git
   cd gemi-chat

2. Install dependencies (using uv for speed):

   uv sync

3. Set up environment:

   • Create a .env file: GOOGLE_API_KEY=your_gemini_api_key
   • Or set the env var: $env:GOOGLE_API_KEY = "your_key"

🛠️ Usage

• Load and index data:

  python -m src.cli --load-data

• Start interactive chat:

  python -m src.cli --chat

• View help:

  python -m src.cli --help

Example chat: Ask questions like "What were UBER's revenue trends in 2020?" and get AI-powered answers based on the filings.

🏗️ Architecture

Auto-Generated Diagrams

Keep architecture diagrams up-to-date with your codebase:

python generate_architecture.py

This script automatically analyzes the src/ directory and regenerates:

• architecture_diagram.md - Detailed component diagram
• codebase_analysis.md - Module analysis report

📁 Project Structure

Gemi_Chat/
├── src/                          # Main package
│   ├── __init__.py              # Package init
│   ├── config.py                # Settings & env vars
│   ├── data_loader.py           # Loads UBER HTML data
│   ├── index_manager.py         # Manages vector indices
│   ├── ageny.py                 # AI agent & tools
│   ├── cli.py                   # Command-line interface
│   ├── custom_console.py        # Console utilities
│   └── google_llm_init.py       # Gemini LLM setup
├── pyproject.toml               # Project config & deps
├── system_prompt.txt            # Agent system prompt
├── .env                         # Environment variables
├── data/UBER/                   # UBER SEC filings
└── storage/                     # Persisted indices

Module Details

• config.py: Centralized configuration (years, paths, API keys).
• data_loader.py: Data ingestion with UnstructuredReader.
• index_manager.py: Vector index creation/persistence.
• ageny.py: Agent setup with query engines and chat loop.
• cli.py: CLI with argparse for commands.
• custom_console.py: Spinners, colors, timers.
• google_llm_init.py: Google Gemini LLM initialization.

🏗️ Architecture

The CLI_Chat application follows a modular RAG (Retrieval-Augmented Generation) architecture:

System Overview

CLI Interface → Data Processing → Vector Indexing → AI Agent → Chat Interface

Core Components Flow

1. Data Pipeline: HTML SEC filings → Document parsing → Vector embeddings → Persistent storage
2. Query Pipeline: User question → Index retrieval → Context augmentation → LLM generation → Response
3. Agent System: FunctionAgent with specialized tools for multi-year financial analysis

Architecture Diagrams

• Detailed Component Diagram: Complete system architecture with all modules and dependencies
• Process Flow Diagram: High-level user journey and data flow

Key Technologies

• LlamaIndex: Vector indexing, query engines, and agent framework
• Google Gemini: LLM for generation and embeddings for semantic search
• Unstructured.io: Document parsing for HTML SEC filings
• RAG Pattern: Retrieval-augmented generation for accurate financial analysis

🤝 Contributing

1. Fork the repo.
2. Create a feature branch.
3. Commit changes.
4. Push and open a PR.

📄 License

MIT License - see LICENSE file for details.

⚠️ Disclaimer

This is for educational purposes only. Not financial advice.