paqPy

Hash directory or file with BLAKE3.

Python bindings to the Rust paq library.

Performance

The Go programming language repository was used as a test data source (157 MB / 14,490 files).

These are upstream paq CLI benchmarks, not measurements of the Python bindings. See the paq 2.0.0 benchmarks documentation for methodology and commands.

Installation

Install From PyPI

pip install paqpy

Install From Repository (Unstable)

Not recommended due to instability of main branch in-between tagged releases.

1. Clone this repository.
2. Run pip install maturin to install maturin.
3. Run maturin develop --release from repository root.

Upgrading to v2

paqpy v2 hashes are incompatible with v1; regenerate stored hashes after upgrading. See release notes.

Usage

import paqpy

from pathlib import Path

source = Path("/path/to/source")
ignore_hidden = True # .dir or .file
source_hash = paqpy.hash_source(source, ignore_hidden, follow_links=False)

print(source_hash)

hash_source(source, ignore_hidden, follow_links=False) returns a hexadecimal hash string. Existing two-argument calls remain supported. By default, symbolic links are hashed by their target-path text, including when the source itself is a symbolic link. Set follow_links=True to hash target contents and traverse linked directories, including targets outside the source tree. Broken links and cycles raise errors when following links.

hash_source uses paq’s fallible Rust API and translates errors into Python exceptions. File system failures raise the corresponding Python OSError subclass, such as FileNotFoundError or PermissionError. Invalid UTF-8 paths raise UnicodeError, invalid source relationships raise ValueError, and other hashing failures raise paqpy.PaqError.

Visit the paq homepage for more details.

License

MIT