python-cvpy is a high-level Python library for working with SAS Viya Computer Vision that offers image and biomedical image analytics. It provides Python-facing abstractions and utilities around SAS image action sets, with emphasis on image retrieval, array conversion, visualization, annotation workflows, and CAS execution utilities.
The package is distributed on PyPI as sas-cvpy and is intended for use alongside SAS Viya and the SAS SWAT Python client.
python-cvpy is designed to make SAS image analytics more accessible from Python by exposing higher-level APIs over CAS-backed image data and related workflows.
The library primarily targets the following usage patterns:
- interacting with SAS Viya image and biomedical image tables
- converting CAS image data tables into Python/numpy-friendly structures
- visualizing 2D and 3D image content using open-source Python tooling
- supporting image annotation-related workflows, including CVAT integration
- assisting with CAS performance tuning for image workloads
To use python-cvpy, you typically need:
- Python 3
- access to a SAS Viya environment
- the
swatPython package - valid SAS backend credentials
Some features also require visualization dependencies such as:
numpypandasmatplotlibmayaviPyQt5
Install from PyPI:
pip install sas-cvpyOr install from source:
git clone https://github.com/sassoftware/python-cvpy.git
cd python-cvpy
pip install .Additional documentation is organized by topic:
- Installation
- Architecture
- API Overview
- Annotation APIs
- Image APIs
- Biomedical Image APIs
- Utilities
- Visualization
Generated API documentation is also available at:
Example material is available in the repository under:
examples/biomedimageexamples/thread_optimization
Contributions are welcome. Please review:
This project is licensed under the Apache License 2.0.