acvl-utils

What it is and what it does

acvl-utils is a collection of utility functions and algorithms maintained by the Applied Computer Vision Lab at Helmholtz Imaging. It provides a dynamic N-dimensional array slicer that adapts to array dimensionality at runtime, avoiding the need to know dimensions beforehand, and a parallel execution wrapper that combines multiprocessing with tqdm progress bars while maintaining result ordering.

The package is built on a substantial scientific computing stack—numpy, torch, SimpleITK, scikit-image, connected-components-3d, batchgenerators, and blosc2—making it most suitable for computer vision and image processing workflows. It is actively maintained and has no known security vulnerabilities, but the heavy dependency footprint means installation can be non-trivial.

Use it for:

• Dynamically slice N-dimensional medical or scientific imaging arrays without knowing dimensionality in advance.
• Run parallel image processing tasks with live progress tracking and ordered results.
• Access ACVL's internal utility implementations for array manipulation in computer vision pipelines.
• Integrate efficient array handling into batch processing workflows for deep learning.
• Prototype array operations in research code where runtime dimensionality is variable.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides utility functions for dynamic N-dimensional array slicing and parallel processing with progress tracking, built on a scientific computing stack.

Yes, if you are working in computer vision or image processing and need dynamic array slicing or parallel processing with progress bars. The heavy dependency footprint means you should verify compatibility with your environment first. No security concerns; actively maintained.

Install

acvl-utils on PyPI

pip install acvl-utils

uv add acvl-utils

poetry add acvl-utils

Installing acvl-utils

Before you install

High install friction due to 7 runtime dependencies including numpy, torch, SimpleITK, and blosc2—a substantial scientific computing stack. Package is actively maintained with recent releases.

License in practice

Licensed under Apache-2.0 (permissive), allowing commercial and private use with minimal restrictions; suitable for most projects.

Quickstart

pip install acvl-utils

from acvl_utils.array_manipulation import Slicer
from acvl_utils.miscellaneous import imap_tqdm

# Dynamic N-dimensional array slicing
slicer = Slicer()

# Parallel map with progress bar
results = imap_tqdm(some_function, iterable)

Requires torch, SimpleITK, and blosc2 to be installed; these are non-trivial compiled dependencies that may require system libraries or build tools.

Verify before relying

• Exact Python version support is unspecified; minimum version requirement unknown.
• Whether the package is suitable for production use or primarily for research/internal ACVL workflows.
• Performance characteristics of the Slicer and imap_tqdm compared to alternatives.

Package facts

Evidence: acvl_utils-0.2.6.tar.gz

Tags

More Scientific/Engineering packages