RAG Chatbot: Get Answers from Documents and Webpages This chatbot, powered by OpenAI's large language model (LLM), can answer your questions based on what you give it! What can it do? • Able to upload documents (PDF, DOCX, TXT) and accept a webpage URL (e.g., Confluence page). • Ask your question and get answers based on the uploaded content.
Getting Started: Clone the repo Perform following commands to run the code in virtual environment • python -m venv ragenv • source ragenv/bin/activate
export OPENAI_API_KEY="your open ai api key"
Install following libraries pip install openai streamlit PyPDF2 docx requests-html langchain exception lxml_html_clean langchain-community chromadb tiktoken
streamlit run chatbot_app.py
Just install the openai library with pip install openai. Known Issues: • Currently, the question field is in the middle of the page ideally should be at the bottom of the page. • Text entered in the question field doesn't automatically clear after submitting. User can edit and ask a new question. Feel free to explore the code and contribute to the project!