tinytrie 0.1.0a6

A minimal type-safe trie (prefix tree) implementation in Python.

TinyTrie

A minimal and type-safe trie (prefix tree) implementation for Python 2+.

Features

• Typed: Works with arbitrary key and value types (Generic[K, V])
• Minimal: Only essential functionalities
• Efficient: Memory-efficient with __slots__
• Iterable: Easily traverse and list all stored sequences
• No external dependencies (except typing on Python <3.5)

Basic Operations

from tinytrie import *

# Create a trie with character (`str`) keys and integer values
root = TrieNode[str, int]()

# Insert some words with values
update(root, 'apple', 1)
update(root, 'app', 2)
update(root, 'banana', 3)
update(root, 'band', 4)

# Search for existing words
assert search(root, 'apple').value == 1
assert search(root, 'app').value == 2
assert search(root, 'banana').value == 3

# Search for non-existent words
assert search(root, 'orange') is None
assert search(root, 'appetizer') is None

update(root, 'apple', 10)
assert search(root, 'apple').value == 10  # Value updated

# Insert a new word
update(root, 'orange', 5)
assert search(root, 'orange').value == 5

# Delete 'apple', 'app' remains
assert delete(root, 'apple') is True
assert search(root, 'apple') is None
assert delete(root, 'apple') is False
assert search(root, 'app') is not None

# Add back 'apple', delete 'app', 'apple' remains
update(root, 'apple', 10)
assert delete(root, 'app') is True
assert search(root, 'app') is None
assert delete(root, 'app') is False
assert search(root, 'apple') is not None

# Try to delete non-existent words
assert delete(root, 'ban') is False
assert delete(root, 'appetizer') is False

# Get common prefix from root (no common prefix)
prefix, _ = longest_common_prefix(root)
assert prefix == []  # No common prefix among all words

# Get common prefix from 'b' subtrie
subtrie_prefix = ['b']
b_subtrie_root = get_subtrie_root(root, subtrie_prefix)
prefix, _ = longest_common_prefix(b_subtrie_root)
assert prefix == ['a', 'n']  # Common between 'banana' and 'band' after 'b'

# Get all words in the trie
words = {''.join(s) for s, _ in collect_sequences(root)}
assert words == {'apple', 'banana', 'band', 'orange'}

# Get all words in the 'ba' subtrie
subtrie_prefix = ['b', 'a']
ba_subtrie_root = get_subtrie_root(root, subtrie_prefix)
words_starting_with_ba = {''.join(s) for s, _ in collect_sequences(ba_subtrie_root, prefix=subtrie_prefix)}
assert words_starting_with_ba == {'banana', 'band'}

Non-String Keys Example

from tinytrie import *

# Create a trie with tuple keys
trajectory_trie = TrieNode[Tuple[int, int], str]()
update(trajectory_trie, [(1, 2), (3, 4)], 'traj1')
update(trajectory_trie, [(1, 2), (5, 6)], 'traj2')

assert search(trajectory_trie, [(1, 2), (3, 4)]).value == 'traj1'
assert search(trajectory_trie, [(1, 2), (5, 6)]).value == 'traj2'
assert search(trajectory_trie, [(1, 2)]) is None  # Partial path

prefix, _ = longest_common_prefix(trajectory_trie)
assert prefix == [(1, 2)]

API Reference

Function	Purpose	Time Complexity
traverse(root: TrieNode[K, V], path: Iterable[K]) -> Iterator[Tuple[Optional[TrieNode[K, V]], K]]	Yields nodes and keys along a path of keys (even if it diverges)	O(n)
get_subtrie_root(root: TrieNode[K, V], path: Iterable[K]) -> Optional[TrieNode[K, V]]	Gets the root node of a subtrie at the end of a path of keys if it exists	O(n)
search(root: TrieNode[K, V], sequence: Iterable[K]) -> Optional[TrieNode[K, V]]	Returns terminal node if sequence is stored in the trie	O(n)
update(root: TrieNode[K, V], sequence: Iterable[K], value: Optional[V] = None) -> TrieNode[K, V]	Inserts or updates a sequence and sets its value	O(n)
delete(root: TrieNode[K, V], sequence: Sequence[K]) -> bool	Removes a sequence and prunes dead nodes	O(n)
longest_common_prefix(root: TrieNode[K, V]) -> Tuple[Sequence[K], TrieNode[K, V]]	Finds the longest common prefix of all sequences and its terminal node	O(m)
collect_sequences(root: TrieNode[K, V], prefix: Optional[List[K]] = None) -> Iterator[Tuple[List[K], TrieNode[K, V]]]	Yields all stored sequences and their terminal nodes	O(n) per sequence

Contributing

Contributions are welcome! Please submit pull requests or open issues on the GitHub repository.

License

This project is licensed under the MIT License.