Spaces:
Running
Running
| title: 06-knowledge-graph | |
| app_file: interface.py | |
| sdk: gradio | |
| sdk_version: 4.44.1 | |
| # Knowledge Graph Project README | |
| ## Overview | |
| This project creates an interactive knowledge graph visualization from RSS feeds. It extracts entities and relationships from news articles using NLP and LLM-based techniques, then visualizes the connections in an interactive graph. | |
| ## Features | |
| - Fetches and aggregates content from multiple RSS news feeds | |
| - Processes text using LLM-based knowledge extraction | |
| - Builds a directed graph of entities and their relationships | |
| - Provides an interactive visualization using Plotly | |
| - Allows selection of specific news sources | |
| ## Installation | |
| 1. Clone the repository | |
| 2. Install the required dependencies: | |
| ```bash | |
| pip install -r requirements.txt | |
| ``` | |
| 3. Download the required spaCy model: | |
| ```bash | |
| python -m spacy download en_core_web_sm | |
| ``` | |
| ## Usage | |
| Run the Gradio interface: | |
| ```bash | |
| python interface.py | |
| ``` | |
| The interface allows you to: | |
| 1. Select which news sources to include | |
| 2. Generate a knowledge graph from the selected sources | |
| 3. View the aggregated feed content and interactive graph visualization | |
| ## Project Structure | |
| - `interface.py`: Main Gradio application with the UI and visualization logic | |
| - `fetch.py`: Functions for retrieving and parsing RSS feeds | |
| - `sources.py`: List of available RSS feed URLs | |
| - `requirements.txt`: Required Python packages | |
| - `tutorials/`: Example notebooks showing the knowledge graph extraction process | |
| ## How It Works | |
| 1. The application fetches recent articles from selected RSS feeds | |
| 2. Content is processed and split into manageable chunks | |
| 3. An LLM (GPT-4o) extracts entities and relationships from the text | |
| 4. A directed graph is constructed from these relationships | |
| 5. The graph is visualized using Plotly with interactive features | |
| ## Dependencies | |
| - spaCy: For NLP processing | |
| - Gradio: For the web interface | |
| - NetworkX: For graph data structures | |
| - Plotly: For interactive visualizations | |
| - LangChain: For LLM-based graph transformations | |
| - OpenAI API: Powers the LLM graph transformer | |
| ## Acknowledgements | |
| This project was inspired by techniques from: | |
| - [Analytics Vidhya: How to Build Knowledge Graph from Text using spaCy](https://www.analyticsvidhya.com/blog/2019/10/how-to-build-knowledge-graph-text-using-spacy/) | |
| - [DataCamp: Knowledge Graph RAG Tutorial](https://www.datacamp.com/tutorial/knowledge-graph-rag) | |
| ## Part of 30 Agents in 30 Days | |
| This project is #6 in the **30 Agents in 30 Days** series, which provides practical AI agent workflows for different stages of product development including research, development, testing, marketing, and sales. |