Personal Portfolio Helper

A conversational AI assistant that answers questions about its creator using RAG (Retrieval Augmented Generation) technology. The bot processes PDF documents containing personal and professional information, creating a knowledge base to provide accurate, context-aware responses about the portfolio owner.

Features

• PDF document processing and vectorization
• Context-aware question answering
• Chat history management
• Professional response formatting
• Integration with OpenAI and Pinecone services

Prerequisites

• Python 3.8+
• OpenAI API key
• Pinecone account and API key
• PDF documents containing portfolio information

Installation

1. Clone the repository:

git clone <your-repository-url>
cd personal-portfolio-helper

2. Install required dependencies:

pip install -r requirements.txt

3. Create a .env file in the root directory with the following variables:

OPENAI_API_KEY=your_openai_api_key
INDEX_NAME=your_pinecone_index_name
PINECONE_API_KEY=your_pinecone_api_key

Project Structure

personal-portfolio-helper/
├── bot.py              # Main chatbot implementation
├── reader.py           # PDF processing and vector database setup
├── requirements.txt    # Project dependencies
├── data/              # Directory for PDF documents
│   └── rag-training-doc.pdf
└── README.md

Components

reader.py

Handles document processing and vector database setup:

• Loads PDF documents using PyPDFLoader
• Splits text into manageable chunks
• Creates embeddings using OpenAI
• Stores vectors in Pinecone database

bot.py

Implements the conversational interface:

• Manages chat history
• Processes user queries with context awareness
• Retrieves relevant information from vector storage
• Generates appropriate responses using OpenAI's GPT model

Usage

1. First, process your PDF documents to create the vector database:

python reader.py

2. Start the chatbot:

python bot.py

3. Interact with the bot by typing questions. Type 'exit' to end the session.

Technical Details

The project uses several key technologies:

• LangChain for RAG implementation
• OpenAI's embeddings and chat models
• Pinecone for vector storage
• Python's dotenv for environment management

The system follows a two-step process:

1. Document Processing: Converting PDF content into searchable vectors
2. Interactive QA: Using chat history and context-aware retrieval for accurate responses

Contributing

Feel free to submit issues and enhancement requests!

Author

Albert Derevski