AI Agent - Intelligent Automation Platform

A comprehensive AI-powered automation platform that combines natural language processing, document analysis, log analysis, calendar management, and GitHub PR review capabilities using LLaMA 3 and LangChain.

🚀 Overview

This AI-assistant is designed to facilitate efficient machine learning model training (RAG-based, storing issues and resolutions in a local vector database), automate pull request reviews with detailed comments on code quality and optimization, manage Gmail calendar, compose and send/delete mail with voice-controlled instructions, and summarize logs/documents/reports.

It uses Python (as the backend) wrapped in FastAPI and React as the Frontend.

👁️ Preview

Features

🤖 AI Chat Agent

Natural language interface powered by LLaMA 3 (Ollama)
Context-aware conversations
Mathematical computation support
Document Q&A capabilities

📄 Document Analysis

Upload and analyze PDF and text documents
AI-powered document summarization
Question-answering on uploaded documents

📊 RAG-based Log Analyzer

Upload and index log files
Semantic search across logs using FAISS vector store
Natural language queries for log analysis
Support for multiple log files with consolidated indexing

📅 Google Calendar Integration

OAuth 2.0 authentication
Create, view, and delete calendar events
Schedule meetings with Google Meet links
Retrieve events by date
Support for attendees and locations

📧 Gmail Integration

Read and summarize emails
Mark emails as read/unread
AI-powered email summarization

🔍 GitHub PR Review

Automated pull request review
Code quality analysis
Vulnerability detection
Best practices suggestions
AI-generated review comments

🧠 RAG based Training & Knowledge Base

Save issue-resolution pairs
RAG based engine selects three most relevant solutions from the local saved resolution history, before sending to the LLM for final verdict.
Training history management

Architecture

AI-python/
├── main_fastapi.py           # Main FastAPI backend server
├── math_ai_agent_doc.py       # LLaMA 3 agent with tool support
├── rag_log_analyzer.py        # RAG-based log analysis
├── training_store.py          # Issue resolution training store
├── pr_review.py               # GitHub PR review automation
├── gmail_auth.py              # Gmail OAuth authentication
├── ai-agent-ui/               # React frontend
├── logs/                      # Log files directory
├── embeddings/                # FAISS vector store
├── uploaded_docs/             # Uploaded documents storage
└── credentials_*.json         # OAuth credentials

Prerequisites

System Requirements

Python 3.10+
Node.js 16+ (for frontend)
Ollama with LLaMA 3 model installed

Ollama Setup

# Install Ollama (if not already installed)
curl -fsSL https://ollama.ai/install.sh | sh

# Pull LLaMA 3 model
ollama pull llama3

# Pull embedding model for RAG
ollama pull nomic-embed-text

Installation

Quick Setup

Use the automated setup script:

./setup.sh

Or follow manual installation:

Backend Setup

Clone the repository:

git clone https://github.com/sandeepknd/AI-agents.git
cd AI-agents

Install Python dependencies:

pip install -r requirements.txt

Set up environment variables:

# Create .env file
echo "GITHUB_TOKEN=your_github_token_here" > .env

Configure Google OAuth:
- Go to Google Cloud Console
- Create a project and enable Google Calendar API and Gmail API
- Create OAuth 2.0 credentials
- Download credentials and save as:
  - credentials_calendar.json (for Calendar)
  - credentials_per_gmail.json (for Gmail)

Frontend Setup

cd ai-agent-ui
npm install

Running the Application

Start Backend Server

# From the root directory
uvicorn main_fastapi:app --reload --host 0.0.0.0 --port 8000

The API will be available at:

API: http://localhost:8000
API Documentation: http://localhost:8000/docs

Start Frontend

# From ai-agent-ui directory
cd ai-agent-ui
npm start

The UI will be available at: http://localhost:3000

API Endpoints

Chat & AI Agent

POST /ask - Send queries to the AI agent

{
  "query": "What is the derivative of x^2?"
}

Document Management

POST /upload - Upload and analyze documents
POST /upload-log - Upload log files for indexing
POST /analyze-log - Query log files

Calendar Operations

GET /authorize-calendar - Start OAuth flow
GET /oauth2callback - OAuth callback
GET /get-events - Get upcoming events
POST /create-event - Create calendar event
GET /get-events-by-date?date=YYYY-MM-DD - Get events for specific date

POST /schedule-meeting - Schedule meeting with Google Meet

{
  "title": "Team Standup",
  "start_time": "2024-03-15T10:00:00",
  "end_time": "2024-03-15T11:00:00",
  "description": "Daily standup meeting",
  "attendees": ["user@example.com"],
  "create_meet_link": true
}

DELETE /delete-event?event_id=xxx - Delete event

Training & Knowledge Base

POST /train-model - Save issue-resolution pair

{
  "issue": "Server not responding",
  "resolution": "Restart the service using systemctl restart app"
}

POST /suggest-resolution - Get AI-powered resolution suggestions
GET /get-training-history - View training history
DELETE /clear-training-history - Clear training data

GitHub PR Review

POST /webhook - GitHub webhook for PR events

POST /comment - Post comment on PR

{
  "pr_url": "https://github.com/user/repo/pull/123",
  "comment": "LGTM! Great work on this feature."
}

POST /generate-comment - Generate AI review comment

{
  "pr_url": "https://github.com/user/repo/pull/123"
}

Usage Examples

Chat with AI Agent

curl -X POST http://localhost:8000/ask \
  -H "Content-Type: application/json" \
  -d '{"query": "What is 15 factorial?"}'

Upload and Analyze Log

curl -X POST http://localhost:8000/upload-log \
  -F "file=@application.log"

curl -X POST http://localhost:8000/analyze-log \
  -H "Content-Type: application/json" \
  -d '{"query": "Show me all error messages"}'

Schedule a Meeting

curl -X POST http://localhost:8000/schedule-meeting \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Project Review",
    "start_time": "2024-03-20T14:00:00",
    "end_time": "2024-03-20T15:00:00",
    "attendees": ["team@example.com"],
    "create_meet_link": true
  }'

Generate PR Review

curl -X POST http://localhost:8000/generate-comment \
  -H "Content-Type: application/json" \
  -d '{"pr_url": "https://github.com/owner/repo/pull/42"}'

Technology Stack

Backend

FastAPI - Modern web framework for APIs
LangChain - LLM orchestration framework
Ollama - Local LLM runtime (LLaMA 3)
FAISS - Vector similarity search
Sentence Transformers - Text embeddings
Google API Client - Calendar and Gmail integration

Frontend

React 19 - UI framework
Axios - HTTP client
Tailwind CSS - Styling
Framer Motion - Animations
React Icons - Icon library

AI/ML

LLaMA 3 - Large language model
nomic-embed-text - Embedding model for RAG
all-MiniLM-L6-v2 - Sentence similarity model

Configuration

Environment Variables

GITHUB_TOKEN=your_github_personal_access_token
HF_TOKEN=your_huggingface_token  # Optional, for higher rate limits

Calendar OAuth Scopes

The application requests the following Google Calendar scopes:

https://www.googleapis.com/auth/calendar

Gmail OAuth Scopes

For Gmail integration:

https://www.googleapis.com/auth/gmail.modify
https://www.googleapis.com/auth/gmail.readonly

Troubleshooting

Ollama Connection Issues

# Check if Ollama is running
ollama list

# Restart Ollama service
systemctl restart ollama  # Linux
# or
brew services restart ollama  # macOS

Import Errors

If you encounter LangChain import errors, ensure you have the correct packages:

pip install --upgrade langchain langchain-community langchain-classic langchain-text-splitters

Port Already in Use

# Find process using port 8000
lsof -i :8000

# Kill the process
kill -9 <PID>

Development

Running Tests

# Frontend tests
cd ai-agent-ui
npm test

# Backend tests (if implemented)
pytest

Code Style

The project follows PEP 8 for Python and Prettier for JavaScript/React.

Security Considerations

API Keys: Never commit credentials_*.json, token.json, or .env files
OAuth Tokens: Tokens are stored locally in token.json and token_mail.pickle
GitHub Token: Store in environment variable or .env file
CORS: Frontend is restricted to http://localhost:3000 in production

Contributing

Fork the repository
Create a feature branch
Make your changes
Submit a pull request

License

[Specify your license here]

Acknowledgments

LangChain for the excellent LLM framework
Ollama for local LLM deployment
FastAPI for the modern Python web framework
React community for the frontend ecosystem

Support

For issues, questions, or contributions, please open an issue on GitHub.

Built with ❤️ using LLaMA 3, LangChain, and FastAPI

Name		Name	Last commit message	Last commit date
Latest commit History 39 Commits
ai-agent-ui		ai-agent-ui
logs		logs
uploaded_docs		uploaded_docs
.gitignore		.gitignore
README.md		README.md
gmail_auth.py		gmail_auth.py
main_fastapi.py		main_fastapi.py
math_ai_agent_doc.py		math_ai_agent_doc.py
pr_review.py		pr_review.py
rag_log_analyzer.py		rag_log_analyzer.py
requirements.txt		requirements.txt
setup.sh		setup.sh
training_store.py		training_store.py

sandeepknd/AI-agents

Folders and files

Latest commit

History

Repository files navigation