hi-tech-AI/help-scout-assistant-using-Pinecone-vector-database

Help Scout Assistant is a document processing and query-response system that leverages Pinecone for vector storage and retrieval. The tool allows you to load PDF documents into a vector store, where they can be queried using OpenAI's language models.

Help Scout Assistant

Overview

Help Scout Assistant is a document processing and query-response system that leverages Pinecone for vector storage and retrieval. The tool allows you to load PDF documents into a vector store, where they can be queried using OpenAI's language models. The retrieval process aims to simulate a customer support assistant by fetching relevant information from the document database in response to user queries.

Features

• PDF Document Processing: Extracts text from PDFs and splits it into manageable chunks.
• Vector Storage with Pinecone: Stores text chunks in Pinecone for efficient retrieval.
• Intelligent Query Handling: Utilizes OpenAI's language model to generate human-like responses to user queries.
• Customizable Settings: The tool can be tailored through various environment variables for API keys and service configurations.

Installation

Prerequisites

• Python 3.7+
• A Pinecone account and API key
• An OpenAI account and API key
• Required Python packages as listed in requirements.txt

Environment Setup

1. Clone Repository

   git clone https://github.com/hi-tech-AI/help-scout-assistant-using-Pinecone-vector-database.git
   cd help-scout-assistant-using-Pinecone-vector-database

2. Install Requirements

   pip install -r requirements.txt

3. Environment Variables

   Set up your environment variables using a .env file:

   PINECONE_API_KEY=your_pinecone_api_key
   PINECONE_CLOUD=aws
   PINECONE_REGION=us-east-1
   OPENAI_API_KEY=your_openai_api_key
   

Usage

The primary interaction with this system happens through the main function, which coordinates the entire workflow of loading documents, creating/updating the index, and handling user queries.

from your_script import main

user_query = "Your question here"
response = main(user_query)
print(response)

Structure

• extract_text_from_pdf: Loads and splits the text from a given PDF path.
• get_index_from_pinecone: Ensures the specified index exists in Pinecone.
• upsert_doc_to_pinecone: Inserts or updates the document in the Pinecone vector store.
• search_query: Executes a search query on the document store and returns the response.
• main: Orchestrates the flow from PDF extraction to query response.

Customization

You can modify the tool’s behavior by adjusting parameters in functions like CharacterTextSplitter, ChatOpenAI, and changing environment variable values in .env.

Contributing

Feel free to open issues or submit pull requests to discuss and improve the functionality.