fast-hash-utils 0.2.2

Fast deterministic dict hashing via mypyc

hash-utils

Fast deterministic dict hashing via mypyc.

Functions

• dict_hash(d) — deterministic hash of a nested dict's full content (keys + values)
• shape_hash(d) — structural hash that ignores string/int/float values, only hashing keys, value types, bools, and container lengths

Install

pip install fast-hash-utils

Usage

from hash_utils import dict_hash, shape_hash

d1 = {"name": "alice", "config": {"enabled": True, "tags": []}}
d2 = {"name": "bob",   "config": {"enabled": True, "tags": []}}

# Full content hash — different names produce different hashes
dict_hash(d1) != dict_hash(d2)

# Shape hash — same structure produces same hash
shape_hash(d1) == shape_hash(d2)

Why

shape_hash enables massive deduplication for jsonschema validation. If 13,000 dicts share the same structure but differ only in string values, they collapse to 1 unique shape — skip 12,999 redundant validations.

Performance

Compiled via mypyc to native C. ~335K ops/s for nested dicts on a single core.

Method	ops/s	Deterministic
shape_hash (mypyc)	335K	Yes
dict_hash (mypyc)	334K	Yes
json.dumps + hash	112K	Yes

Development

python -m venv .venv && source .venv/bin/activate
make install    # editable install with dev deps
make test       # run tests (pure Python or compiled)
make style      # ruff check --fix + format
make clean      # remove build artifacts

License

MIT