Algorithms Library

A comprehensive, zero-dependency collection of common algorithms for Python. The package is designed to be production-ready, strongly typed, and easy to use.

Installation

pip install algorithms-lib

Quick Start

from algorithms_lib import quick_sort, merge_sort, binary_search, bfs, dijkstra

# Sorting
print(quick_sort([3, 1, 4, 1, 5, 9, 2, 6]))
# [1, 1, 2, 3, 4, 5, 6, 9]

# Searching
print(binary_search([1, 3, 5, 7, 9], 5))  # 2

# Graph traversal
graph = {'a': ['b', 'c'], 'b': ['d'], 'c': [], 'd': []}
print(bfs(graph, 'a'))  # ['a', 'b', 'c', 'd']

# Shortest path
weighted = {
    'a': [('b', 1), ('c', 4)],
    'b': [('c', 2), ('d', 5)],
    'c': [('d', 1)],
    'd': [],
}
distances, _ = dijkstra(weighted, 'a')
print(distances['d'])  # 4

Features

• Zero runtime dependencies: No external packages required
• Comprehensive coverage: Sorting, searching, graphs, dynamic programming, strings, greedy, and divide-and-conquer
• Type safe: Includes py.typed marker for type checkers
• Well tested: Full test coverage for normal use, edge cases, and invalid operations
• Clear error messages: Explicit, helpful exceptions

Supported Algorithms

Sorting

• quick_sort: in-place-memory quick sort
• merge_sort: stable merge sort
• heap_sort: binary max-heap sort
• radix_sort: LSD radix sort for non-negative integers
• tim_sort: Python's Timsort wrapper

Searching

• binary_search: classic binary search
• interpolation_search: interpolation search for numeric data
• jump_search: jump search with block size sqrt(n)

Graphs

• bfs: breadth-first search
• dfs: iterative depth-first search
• dijkstra: Dijkstra shortest paths (non-negative weights)
• a_star: A* shortest path with heuristic
• bellman_ford: shortest paths with negative-weight cycle detection

Dynamic Programming

• knapsack_01: 0/1 knapsack maximum value
• longest_common_subsequence: LCS length
• edit_distance: Levenshtein distance

String Algorithms

• kmp_search: Knuth-Morris-Pratt pattern matching
• rabin_karp_search: rolling-hash pattern matching
• boyer_moore_search: Boyer-Moore with bad-character rule

Greedy

• activity_selection: maximum compatible activities
• fractional_knapsack: fractional knapsack maximum value
• huffman_coding: optimal prefix codes

Divide and Conquer

• max_subarray: maximum subarray sum
• count_inversions: inversion count via merge sort
• fast_power: exponentiation by squaring

Development

pip install -e ".[dev]"
pytest test_algorithms_lib.py -v

License

MIT License. See LICENSE for details.