Ragger is an open source custom built pipeline which lets you upload pdfs and have you ask questions about that pdf from the AI model. These Pdfs can include anything from Legal Contracts & Policies, Academic & Scientific Papers, Tech Manuals & Documentation , etc.
- The webpage is fully built from scratch using Html, CSS and javascript.
- The is backend is built on python and is hosted on Hugging Face(using docker and uvicorn) you access the Hugging face space from here
The project requires the following libraries:-
- fastapi
- uvicorn 3. python-multipart 4. langchain 5. langchain-community
- langchain-openai 7. langchain-core 8. langchain-text-splitters 9. pypdf 10. faiss-cpu 11. python-dotenv
Quick install
pip install fastapi uvicorn python-multipart langchain langchain-community langchain-openai langchain-core langchain-text-splitters pypdf faiss-cpu python-dotenv
Are you looking forward to contribute i would love that TBH the webpage isnt very beautiful sooo you can make your custom webpage and Showcase it here
Sure! Here are the steps:-
- Download the backend.py file from the repo.
- Use hugging face or similar webpage to host the backend and obtain an public apiurl. i. you can test if the backend is working or not by using the generated url if you are using fastapi then apiurl/docs will let you try the backend first.
- make/download the frontend from the repo and use javascript to link the frontend to the backend.
- After linking both the frontend and backend the pipeline should work properly.
