pytries/marisa-trie

Health v2 · maintenance only

Adoption not part of the score

1123 stars · 98 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A Python library providing static, memory-efficient trie data structures built on the marisa-trie C++ library via Cython bindings. It stores string keys using 50x-100x less memory than a Python dict while offering comparable lookup speed and fast prefix search.

Use cases

• store large dictionaries of strings with minimal memory
• fast prefix search / autocomplete lookups in Python
• memory-efficient string-to-value mapping
• build token dictionaries for NLP pipelines
• serialize compact string sets to disk

When to choose

• you need immutable string key storage far more compact than a dict
• you need prefix-based queries like .prefixes() or common-prefix search
• you want pip-installable Python bindings for marisa-trie

When to avoid

• you need mutable tries with frequent insertions and deletions
• you need a values() method or fast .prefixes() on BytesTrie/RecordTrie
• you need read()/write() with file-like objects (use pickling instead)

Facets

library · maturity active

search-engine serialization developer-tools python cross-platform trie marisa-trie cython memory-efficient prefix-search string-mapping data-structures algorithms natural-language-processing

Member repositories

For agents

markdown · JSON · MCP: product_card(name="pytries/marisa-trie")