Releases: gschramm/parallelproj
Releases · gschramm/parallelproj
v1.10.0
v1.10.0
- add support for numpy>=2.0
- add tests with numpy 2.0 on python 3.9 and 3.12
- remove tox.ini
v1.9.1
- BUGFIX: fix missing devices in BlockGeometry LOR descriptor (needed for pytorch + cuda backend)
v1.9.0
- add functionality to create scanners, LOR descriptors and projectors for scanners consiting of equal "block" modules
- BUGFIX: correct behavior of TOF kernel truncation which was wrong in the case that the tof bin width was >> tof resolution
v.1.8.0
v1.8.0 (March 20, 2024)
- add function to count event multiplicity
- add more examples (e.g. DePierro and LM SPDHG)
- re-organize folder structure and pyproject.toml
- force array-api-compat<1.5 (bug in 1.5.0)
- use array-api-strict instead of numpy.array_api
v1.7.3
1.7.3 (January 26, 2024)
- BUGFIX:
import importlib.util
instead of import importlib
- print banner
- test also on Windows
v1.7.2
1.7.2 (January 26, 2024)
- require python>=3.9
- replace
distuils.spawn
by shutil.which
v1.7.1
1.7.1 (January 19, 2024)
- BUGFIX: correct bug in the "chunking" of TOF sinogram projections in the python interface
v1.7.0
1.7.0 (January 15, 2024)
- update of documentation
- addition of more examples
- addition of high-level classes for RegularPolygonPETScanner and LOR descriptors
v1.6.2
- BUGFIX: correct use of conj() of scalar value to be array api comp.
- BUGFIX: divided by float() to be array api comp.
v1.6.1
- BUGFIX: add sigma as explicit argument in GaussianFilterOperator and convert correctly to numpy/cupy arrays