Talk2PDFs - Website Chatbot

Talk2PDFs is a web application that lets users interact with PDF documents through a chatbot. Users can upload PDFs or provide URLs, and the chatbot will use the extracted content to answer questions.

Technologies Used

Python
Streamlit - For the web interface.
Ollama - Using the llama3.2 model for language processing.
Visual Studio Build Tools - Required for compiling dependencies like ChromaDB.
1. Download and install Visual Studio Build Tools.
2. During installation, make sure to select the C++ build tools workload.
3. After installation, use pip to install ChromaDB

Key Features

PDF Upload/URL Input: Upload PDFs or provide URLs to process and extract text.
Text Extraction: Extracts and processes text from PDFs for interaction.
Chatbot Interaction: Ask questions related to the uploaded PDFs and get responses.
Text Preview: View a snippet of the extracted text before asking questions.
Real-Time Responses: Quickly get answers based on the content of the documents.
Integration with Vector Database: Uses ChromaDB for efficient document retrieval.
Session Memory: The chatbot retains previous interactions during a session for continuity.
Chat History: Keeps a log of the session’s conversations.

Setup

Using setup.py

You can set up the project using the provided setup.py file. This will automatically install the required dependencies listed in requirements.txt.

Make sure you have a requirements.txt file with the necessary packages.
Run the following command to install the package:

Manual Setup

If you prefer to set up manually:

Install Streamlit, Langchain, Langchain Community and ChromaDB:
```
pip install streamlit langchain langchain_community chromadb
```
Running:
```
streamlit run application.py
```
For running Ollama (LLM):
```
ollama run llama3.2
```

For Developers

Virtual Environment Setup:

Create a virtual environment:
```
python -m venv venv
```
Activate the virtual environment:

On Windows:

venv\Scripts\activate

On Windows:

source venv/bin/activate

eshan-sud/talk2pdfs