Release Notes

0.1.5_2.6.0 [2026-03-19]

Enhancements:

Added examples/run-biomed-clip-from-hugging-face.py example script for zero-shot classification of medical images using the BiomedCLIP model from Hugging Face

Maintenance:

Added git to Dockerfile system dependencies
Upgraded pinned urllib3 to 2.6.0 and brotli to 1.2.0 in the system conda environment to address security vulnerabilities
Added urllib3==2.6.0 and brotli==1.2.0 to project dependencies in pyproject.toml
Reformatted CONTRIBUTING.md for improved line length consistency

Maintenance:

Enhancements:

Added PyTorch inference runner gear for executing custom inference scripts on input files
Added support for dynamic installation of Python packages required by inference scripts
Added GPU acceleration support with NVIDIA CUDA capabilities
Added example inference script for CT image segmentation using totalsegmentator
Added FileCurator pattern for managing file processing workflows
Added optional tagging of processed files via tag configuration parameter
Added API key input support for authentication in inference scripts

Documentation:

Added comprehensive README with usage instructions, workflow diagrams, and examples
Added CONTRIBUTING guide with dependency management, linting, testing, and merge request guidelines

Maintenance:

Added Gitlab CI configuration with test automation, security scanning, and build pipeline
Added Dockerfile using pytorch/pytorch:2.6.0-cuda12.6-cudnn9-runtime base image
Added pre-commit hooks for code quality enforcement (ruff, pytest, gearcheck, hadolint)
Added .dockerignore for optimized Docker build context
Added uv for dependency management with virtual environment configuration
Updated setuptools to 78.1.1 to address security vulnerabilities
Added test suite with parser tests and integration test framework
Added non-root user (flywheel) for secure container execution