Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Making read-all default for reading MR acquisitions from file. (#1227)
Warning: the following 36 commits are "inherited" from PR #1207: * implemented a more flexible alternative to TO_BE_IGNORED macro * moved IgnoreMask to optional arguments of set_data etc. * defined ISMRMRD acquisition flags in Gadgetron.py * implemented acquisition ignoring helper class IgnoreMask in Python * provided example of user-set acquisition ignoring in acquisition_data.py * changed default IgnoreMask for reading from file to ~13bffff * have to use int for ignore mask instead of size_t as SWIG rejects int64 * [ci skip] small amendment in cGT_ISMRMRDAcquisitionsFromFile * changed default value for all_ in AcquisitionData.__init__ to True * put debug prints in test_pSynergistic.py * trying reading MR acquisitions with all_=False * trying reading MR acquisitions with all_=False in noncartesian_recon.py * trying literal_eval instead of eval to placate Codacy * small edit in test_pSynergistic.py to trigger GitHub Actions * small corrections in MR/acquisition_data.py * switched to 64-bit acquisition ignore masks * made some white space adjustments * switched to employing MRAcquisitionData attribute ignore_mask * implemented set_ignore_mask method for MR AcquisitionData Python class * covered all acquisitions ignored cases, fixes #1220 * corrected previous commit * got rid of object_mask * attended to the reviewer comments and suggestions * Update build-test.yml remove `-e` from the ctest line preventing unit test to complete. * interfaced MRAcquisitionData::ignore_mask() into Python * implemented setter for smoothing (convolution) kernel size * updated CHANGES.md * replaced C-style casts with reinterpret_cast in cgadgetron.cpp * if this commit builds, the source of build failures is localised * corrected AcquisitionData.ignore_mask in Gadgetron.py * added missing initialisations of MRAcquisitionData.ignore_mask_ in constructors * switched to defalt value False for all_ in AcquisitionData constructor * smoothness->smoothing_iterations, smth_kernel_size->conv_kernel_halfsize * [ci skip] removed some commented-out and unused stuff * added method size() to C++ ImageData class The remaining 4 commits are the only ones that are actually from this PR: * Managed to switch to read-all-acquisitions default for MR AcquisitionData Failures cased by noise calibration data forced ignoring it, which was a poor solution. It turned out that ignoring these data just on the preliminary test run in AcquisitionProcessor chain (to find out whether the chain needs AcquisitionFinishGadget) appears to fix the problem save for grappa2_1rep.h5 an grappa2_6rep.h5 data. Possible reason for failures with the latter data is the absence of the measurementInformation in xml headers. Assuming this is true, we can now use the proper read-all-acquisition default when reading MR acquisition data, fixing grappa2 data to be addressed by a separate issue. * applied the bugfix from fix-mr2nifti * corrected synergistic complex resampling test (wrong input data) * synchronized Reg.py with fix-mr2nifti to avoid possible merge conflicts --------- Co-authored-by: Edoardo Pasca <edo.paskino@gmail.com>
- Loading branch information