ChatBook is an application powered by LangChain and OpenAI, built using Streamlit, enabling users to interact with an LLM-powered chatbot. This bot is designed to handle various tasks related to text processing, PDF handling, image uploading, and question-answering.
- Python 3.7 or later
- Streamlit
- PyPDF2
- LangChain
- OpenAI's GPT-3.5
- MongoDB
- Pinecone
- Additional libraries specified in the
requirements.txtfile
-
Clone the repository:
git clone https://github.com/your_username/ChattyPDF.git cd ChattyPDF -
Install dependencies:
pip install -r requirements.txt
-
Ensure you have the necessary API keys and configurations for the following services:
- MongoDB: Replace
st.secrets["mongo"]and configure MongoDB accordingly. - Pinecone: Replace
st.secrets["pinecone"]and set up your Pinecone account.
- MongoDB: Replace
-
Run the application:
streamlit run chatty_pdf.py
- Upon running the Streamlit app, the sidebar offers different options:
- Upload PDFs: Upload PDF files to process and store their content for querying.
- Partially delete PDF: Remove specific content from a previously uploaded PDF.
- Ask Question: Interact with the chatbot to get answers related to uploaded PDFs.
- Upload Image: Add images along with titles and descriptions for retrieval.
- Search Image: Find uploaded image via keywords.
- Delete File: Remove stored data associated with uploaded PDFs or images.
- Upload PDFs:
- Upload one or multiple PDF files.
- Extract text content from uploaded PDFs and store it in Pinecone index.
- Partially delete PDF:
- Delete specific content from a PDF stored in the index without affecting other information.
- Ask Question:
- Interact with the chatbot to ask questions about the uploaded PDFs.
- Upload Image:
- Upload images with titles and descriptions for retrieval.
- Search Image:
- Search via a keyword for any uploaded image.
- Delete File:
- Delete stored data associated with uploaded PDFs or images.