Metadata-Version: 2.4
Name: tdkg
Version: 0.1.2
Summary: Time Decayed Knowledge Graph for AI memory systems
Author: Siddhant Sukhatankar
License: MIT
Project-URL: Homepage, https://github.com/sidvsukhi/tdkg
Project-URL: Repository, https://github.com/sidvsukhi/tdkg
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.8
Description-Content-Type: text/markdown

# tdkg

**tdkg (Temporal Decaying Knowledge Graph)** is a lightweight Python library for building memory systems that evolve over time.

It is designed for applications where information should not just be stored, but **reinforced through usage and forgotten naturally over time**.

Think of it as a simple cognitive memory layer for AI systems, agents, or personal knowledge graphs.

---

## 🚀 Key Idea

Traditional storage systems treat all data equally.

`tdkg` introduces a different model:

- Frequently accessed information becomes stronger
- Rarely accessed information decays over time
- Memory behaves more like human cognition

---

## 🧠 Core Concept: Node

Each piece of information is stored as a `Node`:

- Has content
- Tracks access frequency
- Tracks last access time
- Has an evolving importance score

---

## 📦 Installation

```bash
pip install tdkg
