Cachehash is a lightweight Python library for caching file, directory, and value data in SQLite using fast xxhash checksums.
It is designed to memoize expensive computations by binding results to the content hash of files or directories, so you can skip recomputation when inputs have not changed.
- ⚡ Fast hashing with xxhash
- 📂 Hash individual files, whole directories (structure + metadata + file content), or arbitrary values
- 💾 Store and retrieve JSON-serializable Python values
- 🗄️ Backed by SQLite, easy to inspect and portable
- 🔑 Get/set by path + hash, or by hash alone
- 🧰 Simple API for integration into data pipelines or caching layers
pip install cachehashRequires Python 3.9 – 3.13.
from pathlib import Path
from cachehash.main import Cache
cache = Cache()
# Hash a file
file_hash = cache.hash_file(Path("data.csv"))
print(file_hash)
# Store computation result tied to the file
cache.set("data.csv", {"rows": 1234, "checksum": file_hash})
# Later, retrieve result if file unchanged
result = cache.get("data.csv")
print(result)hash_file(path)→ str : Hash file contentshash_directory(path)→ str : Hash entire directory treehash_value(value)→ str : Hash an arbitrary string/bytesget(path)→ Any | None : Retrieve value by file path + hashget_by_hash(path)→ Any | None : Retrieve value by hash onlyset(path, value, append=False): Store a value by file path + hashset_by_hash(key, hash, value, append=False): Store a value for precomputed hashset_value(key, value): Store a value by key only (not tied to content)get_value(key): Get a value by key only
Clone the repo and install with dev dependencies:
git clone https://github.com/VerinFast/cachehash.git
cd cachehash
pip install -e .[dev]Run tests:
pytestFormat code:
black srcFree for non-commercial use. See LICENSE for details.