fast-array-utils
Decision gist · record as of 2026-08-14
Yes, if you work with multiple array formats (sparse, GPU, distributed, file-backed) and want a single conversion and stats API. The low install friction, active maintenance, and zero known vulnerabilities make it safe to adopt. Be aware of the MPL-2.0 copyleft license if you redistribute. Verify that optional dependencies align with your environment before relying on GPU or Dask features.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires Python 3.12 or later.
- Optional dependencies (scipy, cupy, dask, h5py, zarr) must be installed separately or via extras to use those array types.
- Low friction: pure Python wheel with two runtime dependencies (array-api-compat and numpy).
License · maintenance · safety
MPL-2.0 (copyleft) — MPL-2.0 (copyleft): you must disclose source code modifications and distribute under the same license if you redistribute the package or derivative works.
last release 2026-07-17 (28 days) · last repo commit 2026-08-14 · 15 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 417,790 downloads/mo, #6,807 on PyPI
Alternatives
Verify before relying
pip install 'fast-array-utils[accel]'
from fast_array_utils.conv import to_dense
from fast_array_utils import stats
numpy_arr = to_dense(sparse_arr)
col_sums = stats.sum(arr_2d, axis=0)- Whether optional dependencies (scipy, cupy, dask, h5py, zarr, anndata) are automatically installed with extras or must be added separately.
- Performance characteristics and memory overhead compared to calling array libraries directly.
- Whether the testing extra includes test fixtures or only testing utilities.
What it is and what it does
fast-array-utils provides a unified interface to convert and analyze arrays across multiple storage and compute backends—NumPy, SciPy sparse matrices, CuPy GPU arrays, Dask distributed arrays, and file-backed formats like HDF5 and Zarr. It abstracts away the differences between these formats so you can densify sparse data, move arrays between CPU and GPU memory, and compute statistics without rewriting logic for each array type.
The package is built on top of numpy and array-api-compat, keeping its core lightweight. It targets modern Python (3.12+) and is marked Production/Stable. Use it when you need to work with heterogeneous array sources in the same pipeline—for instance, reading sparse matrices from disk, densifying them, computing row statistics, and optionally moving results to GPU for downstream processing.
Use it for
- Densify sparse matrices before passing to dense-only algorithms.
- Compute row or column statistics across mixed array types without format-specific code.
- Move Dask or CuPy arrays to CPU memory for serialization or downstream libraries.
- Unify array handling in pipelines that mix dense, sparse, and GPU data.
- Calculate aggregate statistics across distributed or GPU-resident arrays.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you work with multiple array formats (sparse, GPU, distributed, file-backed) and want a single conversion and stats API.
The low install friction, active maintenance, and zero known vulnerabilities make it safe to adopt. Be aware of the MPL-2.0 copyleft license if you redistribute. Verify that optional dependencies align with your environment before relying on GPU or Dask features.
Install
fast-array-utils on PyPI
Before you install
Low friction: pure Python wheel with two runtime dependencies (array-api-compat and numpy). Active maintenance with recent releases; last commit 2026-08-14. Requires Python 3.12 or later.
Requires Python 3.12 or later. Optional dependencies (scipy, cupy, dask, h5py, zarr) must be installed separately or via extras to use those array types.
License in practice
MPL-2.0 (copyleft): you must disclose source code modifications and distribute under the same license if you redistribute the package or derivative works.
Quickstart
pip install 'fast-array-utils[accel]'
from fast_array_utils.conv import to_dense
from fast_array_utils import stats
numpy_arr = to_dense(sparse_arr)
col_sums = stats.sum(arr_2d, axis=0)Verify before relying
- Whether optional dependencies (scipy, cupy, dask, h5py, zarr, anndata) are automatically installed with extras or must be added separately.
- Performance characteristics and memory overhead compared to calling array libraries directly.
- Whether the testing extra includes test fixtures or only testing utilities.
Package facts
| License | MPL-2.0 copyleft |
| Python support | Supports the current Python release >=3.12 |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | 2 packagesarray-api-compatnumpy |
| Maintenance | Actively maintained 28 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 417,790 / month, #6,807 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Development Status :: 5 - Production/StableIntended Audience :: DevelopersProgramming Language :: Python :: 3 :: OnlyProgramming Language :: Python :: 3.12Programming Language :: Python :: 3.13Programming Language :: Python :: 3.14 |
Evidence: fast_array_utils-1.5-py3-none-any.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “sparse array conversion”
- fast-array-utilsConverts and computes statistics across sparse, dense, GPU, and…
- sparseProvides multi-dimensional sparse array data structures compatible…
- tiledbTileDB-Py provides a Python interface to the TileDB array storage…
Give your agent the search over MCP, or paste the wish link into any chat.
More Scientific/Engineering packages
NumPy provides an N-dimensional array object and a comprehensive suite of mathematical, linear algebra, Fourier transform, and random number functions for scientific computing in Python.
pandas provides fast, flexible data structures (Series and DataFrame) for loading, cleaning, transforming, and analyzing labeled or relational data in Python.
scipy provides numerical algorithms for mathematics, science, and engineering—including optimization, integration, linear algebra, Fourier transforms, signal and image processing, and ODE solvers—built on numpy arrays.
scikit-learn provides a comprehensive Python library for supervised and unsupervised machine learning, including classification, regression, clustering, dimensionality reduction, and model evaluation tools built on NumPy and SciPy.
Install it if you need to train, evaluate, or deploy supervised or unsupervised learning models.
dill extends Python's pickle module to serialize and deserialize a much wider range of Python objects, including functions, lambdas, classes, and interpreter sessions, to byte streams for storage or network transmission.
Multiprocess is an enhanced fork of Python's standard multiprocessing library that uses dill for better serialization, allowing you to spawn processes with a threading-like API and share complex objects between them.
Install it if you use multiprocessing and encounter pickle serialization limits with lambdas or complex objects.
See also array-api-compat · sparse · numbagg · tiledb · nvidia-cusparse · nvidia-cusparse-cu12 · numpy-groupies · tabmat · sparse-dot-topn · boost-histogram