-- All credits go to user @thu_vu92 and reddit.com
This application is a simple streamlit based web interface to take in any summarized text data for a show,movie or any other media for which we have a synopsis.
It extracts graph entities(entities & relationships) from text input using Lang graph and OpenAi's gpt4-o, and uses Pyviz to create a visualized representation of knowledge graph, in order to summarize information through a DAG(Directed Acyclic Graph).
- Dual input methods: text upload(.txt) or a direct text input
- Interactive visualization for knowledge graph
- Entity relationship extraction based on ChatGPT's gpt-4o model
- Physics based layout for customizable graph display
You need to have Python>=3.8 installed, and preferable a code editor like VScode.
Download & Run the requirements.txt file using the below command:
pip install -r requirements.txt
You can set it up in a new virtual environment if needed.
- Python>=3.8
- Open AI API keys (Go to Open AI API keys, and create an API key)
You might need to store the secret credentail key from Open AI inside a .env file with the name OPENAI_API_KEY
To run the above app, in your terminal or powershell, just type:
streamlit run app.run
and it will open a new window in your browser to run this app.
Note: If you get some prompts in the command shell for email, etc, just press enter and keep it running.
For closing this app, press Ctrl+C in the terminal to close.