The purpose of this Learning Tool is to serve as an AI for Computer Science Papers, capable of referencing and extracting additional information from OpenAI. The tool displays references to papers and the corresponding response from OpenAI.
- Custom Agent
- Qdrant Vector DB for Storing PDFs
- Custom Tools
- Qdrant Document Loader Agent (inprogress)
- An Integrated Simple Streamlit UI, it has basic memory
This tool is highly inspired by llm_agents, Virat's work & ChattierGPT-UI
-
Navigate to the directory where the repository was downloaded
cd cspaper-ai
-
Install the required dependencies
pip install -r requirements.txt
-
Configure OpenAI Key
- If Using OpenAI key, simply
export OPENAI_API_KEY=*****
- If want to use config file, rename
config_template.ini
->config.ini
file inside thedatabase_agent
dir & update either Azure or OpenAI config
By completing these steps, you have properly configured the API Keys for your project.
- If Using OpenAI key, simply
- Copy all desired PDFs in a directory
- Update the directory path in
./config.py
>>pdf_uploadpath
- Run the command
python -m paperai upload
Its a 1 time job and once, the pdfs are loaded then user can ignore this step for next runs
python -m paperai run
- The Qdrant Vector Database is currently not operating in server mode, therefore document uploads are only permitted once. If new PDFs are to be added, it is necessary to delete the collection using
client.delete_collection(collection_name="test_papers")
and then re-upload the documents.