pygraphkit 0.1.1

A clean and reusable graph algorithms toolkit

graphkit

graphkit is a clean, reusable Python library providing standard graph algorithms with a unified and intuitive API.

It is designed for:

• Learning and revision of graph algorithms
• Interview and competitive programming preparation
• Real-world projects requiring graph processing
• Avoiding repeated reimplementation of well-known algorithms

---

✨ Features

• Simple Graph abstraction
• Object-oriented API
• Readable and canonical implementations
• Fully testable and extensible
• No external dependencies

Algorithms included

• Shortest Path

  • Dijkstra’s Algorithm
  • Bellman-Ford Algorithm

• Minimum Spanning Tree

  • Kruskal’s Algorithm

• Traversals

  • Breadth-First Search (BFS)
  • Depth-First Search (DFS)

---

📦 Installation

pip install pygraphkit

For development:

pip install -e .

---

🚀 Quick Start

from graphkit import Graph

g = Graph()

g.add_edge(1, 2, 4)
g.add_edge(1, 3, 1)
g.add_edge(3, 2, 2)

print(g.dijkstra(1))

Output

{1: 0, 2: 3, 3: 1}

---

🧠 Core Concept

Graph Abstraction

All algorithms operate on a single Graph class.

Graph(directed=False)

• directed=False → undirected graph
• directed=True → directed graph

Adding edges

g.add_edge(u, v, weight)

• Default weight is 1
• For undirected graphs, edges are added both ways

---

📐 API Reference

Dijkstra’s Algorithm

Finds shortest paths from a source node (non-negative weights).

g.dijkstra(source)

Returns:

{node: shortest_distance}

---

Bellman-Ford Algorithm

Supports negative edge weights and detects negative cycles.

g.bellman_ford(source)

Raises:

ValueError: Negative cycle detected

---

Kruskal’s Algorithm

Computes Minimum Spanning Tree (undirected graphs only).

mst, total_weight = g.kruskal()

Returns:

• mst: list of edges (u, v, w)
• total_weight: sum of MST edge weights

---

Breadth-First Search (BFS)

g.bfs(source)

Returns traversal order as a list.

---

Depth-First Search (DFS)

g.dfs(source)

Returns traversal order as a list.

---

🧪 Testing

graphkit uses pytest for testing all core algorithms.

The test suite covers:

• Shortest path correctness
• Negative edge weights
• Negative cycle detection
• Disconnected graphs
• Error handling for invalid usage

Run tests locally:

pip install -e .
pytest -v

All tests must pass before a release is published.

📁 Project Structure

graphkit/
├── graphkit/
│   ├── graph.py
│   ├── algorithms/
│   ├── utils/
│   └── __init__.py
│
├── tests/
│   └── test_*.py
│
├── README.md
├── pyproject.toml
└── LICENSE

---

🎯 Design Philosophy

• One canonical implementation per algorithm
• No premature optimization
• Code clarity over cleverness
• Easy to rewrite during competitive programming
• Reusable in real-world systems

---

🛣️ Roadmap

Planned additions:

• Prim’s Algorithm
• Floyd-Warshall Algorithm
• Topological Sort
• Strongly Connected Components
• Benchmarking utilities

---

🤝 Contributing

Contributions are welcome.

• Add algorithms
• Improve tests
• Improve documentation

Please keep implementations:

• Clean
• Readable
• Well-tested

---

📜 License

MIT License

---

👤 Author

Gaurav Khakse

If you use this library for learning or projects, a ⭐ on GitHub is appreciated.