README.md

metadata

title: Knowledge Mapper
emoji: 🐢
colorFrom: green
colorTo: pink
sdk: gradio
sdk_version: 5.33.1
app_file: app.py
pinned: false
short_description: AI agent that transforms a text to knowledge map

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

🧠 AI Knowledge Map Generator

This project is an advanced AI-powered application that transforms unstructured text, topics, or questions into structured, interactive knowledge graphs. It leverages a agent system built with LangChain to analyze input, extract key concepts, identify relationships, and generate both a visual map and a textual summary, providing a powerful tool for learning and research.

Tags: agent-demo-track

---

✨ Features

• Multi-Source Input: Accepts direct text, a topic to search on Wikipedia, or a question requiring a web search.
• Agentic Workflow: Uses a system of AI agents to route requests, extract concepts, find relationships, and summarize the findings.
• Interactive Visualization: Generates a dynamic and interactive knowledge graph using Pyvis, where you can move nodes and explore connections.
• AI-Generated Summaries: Provides a coherent, LLM-generated summary that explains the concepts and their relationships in plain language.
• Multi-LLM Support: Easily switch between various LLM providers, including Gemini, OpenAI, Groq, Anthropic, and more.
• Downloadable Output: Allows you to download the generated knowledge graph as a self-contained HTML file.

---

🛠️ How It Works

1. Router Agent: First the agent analyzes your input. It decides if the text is self-contained, if it needs to perform a Wikipedia search for a general topic, or if it requires a real-time web search.
2. Extraction Agent: After gathering the necessary text, the agent reads it and extracts the most important concepts, assigning each an importance score.
3. Linking Agent: The agent analyzes the extracted concepts and the source text to identify all the meaningful relationships between them, describing each link (e.g., "is a type of," "enables," "is composed of").
4. Summarizer Agent: Finally, the agent takes the list of concepts and their relationships and generates a concise, human-readable summary of the entire knowledge graph.
5. Rendering: The system uses the extracted data to render the interactive pyvis graph and display the summary in the Gradio interface.

---

🚀 Video Overview

Watch this short video for a complete overview of the app's purpose, features, and a demonstration of how to use it.

• Video Link: Knowledge Map Generator

---

⚙️ Setup and Configuration

To run this project locally, you'll need to install the required Python packages and configure your environment variables.

1. Software Dependencies

All the necessary Python libraries are listed in the requirements.txt file. You can install them all with a single command:

pip install -r requirements.txt

2. Environment Variables

This application requires API keys to connect to services. The recommended way to manage these is to create a .env file in the root directory of your project.

Create a file named .env and add the following content.

# ---------------- LangSmith Tracing (Recommended for Debugging) ----------------
# Set this to 'true' to enable detailed tracing of your agent's execution in LangSmith.
LANGCHAIN_TRACING_V2="true"
# Your LangSmith API key, available from the LangSmith settings page.
LANGCHAIN_API_KEY="YOUR_LANGSMITH_API_KEY_HERE"


# ---------------- Required Tool APIs ----------------
# The agent uses Tavily for its web search tool.
TAVILY_API_KEY="YOUR_TAVILY_API_KEY_HERE"