| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
| |
* added dot product between datacontainers
closes #208
implements dot product by flattening the data in a vector and calculating
the inner product on the vectors
added unittest
* use more efficient ravel than flatten
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* initial revision
* Removed class members of Algorithm class
added update_objective
* initial version. Fix inline __idiv__
* First implementation of CompositeOperator/DataContainer
* removed __getitem__ added get_item added shape
* added CGLS
* working unit test, initial tomography test
* added reverse multiplication of operator with number
* added operators directory
* fixed typo
* added unittest for CompositeDataContainer
* fix TomoIdentity with scalar
* check numerical types from numpy
* add default stop criterion and run method
* add run method
* first working implementation of CGLS with CompositeOperator/DataContainer
notice problem with _rmul_ and _mul_ methods precedence with numpy.
* removed line endings
* removed dos line ending
* Added allocate to ImageGeometry and AcquisitionGeometry
* remove composite operator
* Delete Algorithms.py
* remove setup.py from PR
* readded setup.py
* added newline at the end of the file
* added newline at EOF in setup.py
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* initial revision
* Removed class members of Algorithm class
added update_objective
* initial version. Fix inline __idiv__
* First implementation of CompositeOperator/DataContainer
* removed __getitem__ added get_item added shape
* added CGLS
* working unit test, initial tomography test
* added reverse multiplication of operator with number
* added operators directory
* fixed typo
* added unittest for CompositeDataContainer
* fix TomoIdentity with scalar
* check numerical types from numpy
* add default stop criterion and run method
* add run method
* first working implementation of CGLS with CompositeOperator/DataContainer
notice problem with _rmul_ and _mul_ methods precedence with numpy.
* new Algorithm class and algorithms in separate files
Added new Algorithm class and derivatives in different files for
GradientDescent, CGLS, FBPD, FISTA
* added algorithms and restored CIL_VERSION env variable
* removed Algorithms.py
* modified run and renamed a few members/methods
* uses squared_norm
* renamed get_current_objective to get_last_objective
update_objective can be issued every N iteration, default 1.
fixed run method to run N iterations within the stop criterion.
* load class as module files
* force py line endings to LF
* updates
* call super __init__ as first thing
* unit tests are now to be found in test directory
unit tests are now split in several files in the directory test
* install algorithms module
* Implementation with Algorithm
* skip Reader tests
* unittest for linux
* commented not needed import Iterable
* removed explicit return from __init__
* remove composite operator file
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* added squared_norm
closes #203
* added norm and squared_norm
closes #203
* Power method uses norm, bugfixes
* fix power method
|
| |
|
|\
| |
| | |
Colourbay initial demo
|
| |\
| |/
|/| |
|
| | |
|
|\ \
| | |
| | | |
Universal build script
|
| |\ \
| |/ /
|/| | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* UPDATE: CIL VERSION needs to be set
* UPDATE: jenkins build and variants
* UPDATE: channels conda-forge and ccpi
|
| | | |
|
| | | |
|
| |\ \ |
|
| | | |
| | | |
| | | | |
ADD: build status from jenkins embeddable build status plugin
|
| | | | |
|
| | | | |
|
| | | | |
|
| | | | |
|
| | | | |
|
| | | | |
|
| |/ / |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
|/ / |
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* replaces in with explicit test
closes #139
* add read of single slice or subset
* get_image_keys
* reader reads single slice
* adds get_acquisition_data_slice and get_acquisition_data_subset
it is possible to read in only a subset of the dataset.
* added multifile_nexus.py
* cleaned example
* fix normaliser and reader
* wip
* wip for reader
* added support for non standard?? nexus files
updated example with normalisation.
* fix camelcase
* Added initial unittest for reader
bugfix for python 2.7
* fixes for python 2.7
* remove duplicate test requirement
* minimal unittest for get_acquisition_data_subset
* added SPDHG
* work in progress
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* replaces in with explicit test
closes #139
* add read of single slice or subset
* get_image_keys
* reader reads single slice
* adds get_acquisition_data_slice and get_acquisition_data_subset
it is possible to read in only a subset of the dataset.
* added multifile_nexus.py
* cleaned example
* fix normaliser and reader
* wip
* wip for reader
* added support for non standard?? nexus files
updated example with normalisation.
* fix camelcase
* Added initial unittest for reader
bugfix for python 2.7
* fixes for python 2.7
* remove duplicate test requirement
* minimal unittest for get_acquisition_data_subset
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* enable win build with unittest. skip cvx
* unittest on funcs
* fixed Norm1 and function
* test chaining
* a few fixes for unittest
* investigating Norm2
* restored Norm2
TODO: rewrite it and test
* removed comment
|
| |
| |
| | |
fixed unit test for python2.7
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
closes #148
closes #156
closes #157
closes #158
* split run_test for algorithms
* fix FBPD
adds operator as input argument
changes naming of variables to explicit
* implement proximal for ZeroFun and Norm1
proximal method must be correctly implemented. lead to error #158
* fix unit test and cvx demo
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Added working demo of colourbay data load and recon.
* Added IMAT white-beam demo loading summed fits files
* corrections to normalization and log with zeroes in flats
* IMAT data multichannel script started
* script to reconstruct multi-channel imat data updated
* some updates to demo
|
| |
| |
| | |
addresses this loophole https://github.com/CCPPETMR/Hackathon-SIRF/issues/12#issuecomment-430944064
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* added inplace algebra correctly
* readded object to AcquisitionGeometry and ImageGeometry
* Add more memory management
* added unary and binary operations to DataContainer
* Added unary and binary operations
added also unittest
* added unittest
* added unit tests for binary operators
* bugfixes and remove prints
* dafault initialisation behaviour not to deep_copy
* FISTA to use the efficient memory handling additions
* add calls with input out=None
Starts to address #134
* add methods with arg out=None
* add test for memhandle
* major bugfix after removal of create_image_data
* fixed FISTA memhandle
* some bugfix in FISTA and funcs ops
* debug in progress
* Fixed several bugs in memory optimised code
Added methods allocate_adjoint and allocate_direct to Operator.
* test with cvx and adjustments
* fixed unittest, need to fix inline algebra
* fixed inline algebra definition
* initial cvxpy unittest
* adds cvx demo as unit test
closes #150
check equality of cvx calculated with respect to FISTA and FBDP
|
| |
| |
| |
| |
| |
| | |
* Quick prototype of SIRT with nonnegativity added
* Add indicator function for boxconstrint in SIRT and FISTA with demo
|
| |
| |
| | |
closes #139
|
| | |
|
| | |
|
| |
| |
| | |
closes #126
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|