Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
* added a new method to blobcorrector.correctorclass to do distorsion correction using a pixel LUT * Attempt to manage openmp with multiprocessing * Attempt to manage openmp with multiprocessing * Attempt to manage openmp with multiprocessing * drop 3.7 for 3.8 (but keep 2.7 for now) * Update blobcorrector.py Added new property self.dim with detector dimension to blobcorrector.correctorclass, inferred from the splinefile (xmin, xmax, ymin, ymax) * Update blobcorrector.py * TODO: fix fscan2d to be easier to use * Thanks to Andrew Reader's youtube video * added 3.11 (3.x is busted due to numba) * auto reshape for fscan2d * setting threads only for multiprocessing * fixup fscan2d * fixup fscan2d * Update build_flake_pytest_ubuntu2004.yml * Update build_flake_pytest_ubuntu2004.yml * Update setup.py * Update sigma_3_matrices.py (py2, print) * Update properties.py (py2: print) * Update test_datasets.py (py2: fstring/print) * Update cImageD11.py : use OMP_NUM_THREADS and py27 * Enable manually specifying detector and motor names * matrix vs dcm with old scipy * fix py27 windows * binary wheels for win32/py27 * 'fixup' * 'fixup' * 'fixup' * Update appveyor.yml * Update appveyor.yml * Update appveyor.yml * Update appveyor.yml * Update appveyor.yml * Update appveyor.yml * Update appveyor.yml * Add build wheels and tarball * Update README.md * Update __init__.py * Create eps_sig_solver_NEW.py Create a new solver to compute strain and stress from ImageD11 ubis * Bugfix assemble_label.py set rather than list randomly permuted the order of scans. Oups. * compatible for python2.7 * fix file path * corrected mistakes in eps_sig_solve_NEW.py + updated invariant properties function - corrected a few mistakes and typos in function description - corrected a mistake in convert_tensors_to_vect - changed invariant_prop function to return von Mises stress / strain instead of octahedral shear stress / strain * Update eps_sig_solver_NEW.py * corrected typo in form_stiffness_tensor * updated Cij_symmetry * Update eps_sig_solver_NEW.py * Rename eps_sig_solver_NEW.py to stress.py * Rename ImageD11/eps_sig_solver.py to ImageD11/depreciated/eps_sig_solver.py * Update stress.py * Update stress.py * Update stress.py * Update stress.py * Update stress.py * Add scanning 3DXRD example notebooks * Remove f-string for 2.7 compat. * Exclude Python 3.12 in setup.py Until Numba updates to Python 3.12 This should fix the CI for Python 3.x * Exclude 3.12 from GitHub Actions CI * Fix GitHub actions yml typo * Third time lucky * Use GitHub Actions badges in README.md * Correct typo * Update README.md * fix the log strain - needs testing better * helper function from several notebooks * mlem and centers code * function from notebooks * function from notebooks * Change int32 and float32 to 64 * remove broken test * put stress on hold while updating it * Update __init__.py v2.0.2 * Update README.md v2.0.2 * Update release.yml : no 3.12 version * Allow a fuller mask, to mask powder rings for example * fix mask for powder removal and reformat (black) * Resurrect tests for columnfile.py, correct np.int In the Github Actions CI, the command "python -m pytest" is run to test ImageD11. By default, this only runs test files with the naming convention `test_*.py` Renamed `testcolumnfile.py` to `test_columnfile.py` to fix this. Also, replaced `np.int` references `int` to fix numpy depreciation errors. * Fix np.int64 * Add colfile tests to run_tests.py * Significant notebook layout changes * Parallelised iradon * Ensure 0 fill value for np.interp * Add mask to MLEM * First attempt at enforcing file paths * A few bugfixes * update cibuildwheel and upload|download-artifacts * eps_sig_solver was deprecated * eps_sig_solver was deprecated + black formatted * Python 3.12 support * Add col3d file to dataset paths * Try to get rid of the fork problem * point by point - needs some more docs to expain pars * fixup 2.7 * Add outsize support to MLEM * Clean up comment * convert nnz to ip function * error on typo in group names * fix forking problem/error * get the warnings to print, but only when needed * for CI only having 1 thread? * CI fails. Not sure why. * fixup * set forkserver here, todo: windows? * and for windows that does not fork anyway * and for windows that does not fork anyway * and for windows that does not fork anyway * percent, not fraction * tested on SS316 * was not tested yet. pythonpath * tested once: should not be forking * seen to run once * with a threshold image option * Notebook for par file building from poni * A few small sinogram bugfixes * self.shape is undefined * Significant scanning notebook changes General changes: - Moved quite a few common functions into utils.py. These need properly incorporating into ImageD11 in the future - All notebooks updated to obey DataSet filestructures for peaks and grains files 0_S3DXRD_segment: - Added widget written by Pierre-Olivier for adjusting segmenter parameters - Added slurm bash "listener" that waits for slurm job to complete before next cell completes - Moves individual sparse files to their own subfolder after assembling 1_S3DXRD_index: - More sensible indexer that prioritises low-multiplicity rings - Grains now saved in HDF5 format - Added another example notebook for indexing a minor phase 2_S3DXRD_sinograms: - Added whole-sample sinogram reconstruction to use as a mask for individual grain reconstructions. - Added MLEM refinement stage (uses SLURM) to grain reconstructions using run_mlem_recon.py - Added another example notebook for mapping a minor phase (uses grain recons for centre-of-mass) 3_S3DXRD_strain_maps_pbp: - Performs a point-by-point refinement of the grain-averaged UBI using 2D peak selection - Uses grain labels array from sinograms for grain shapes - Should give "relative" Type-III strains 4_S3DXRD_plot_both_phases: - Small example notebook for plotting both phases in the same map - Some initial work on orientation relationship identification using Orix (not finished yet) * Add box-beam example notebooks - Similar to scanning notebooks but for box-beam 3DXRD - Currently only working for Frelon data - Uses Frelon segmenter from ma4750 (in sandbox) for peaksearching - Global parameter refinement notebook - Uses makemap.py to refine grain positions - Some extra notebooks for importing multiple slices in a z series and looking at peak shapes (still under development) * Build parameter file from CIF and pyFAI PONI * Clear notebook outputs * Correct pyplot typo * Allow flake 2.7 linting to pass * Python 2 compat * Remove duplicate segmentation notebook * Common utils between scanning and box-beam notebooks - Moved a few more functions out of notebooks into utils - This significantly simplifies the notebooks - MLEM recon: y0 now gets written to/read from the grain HDF5 file rather than command-line option * Replace FileNotFoundError with IOError for py2.7 * Minor phase scanning notebooks now use nb_utils, needs testing * Correct ImageD11_gui bugs - broken log buttons and tth_histogram * Rename and add batch scripting * Add batch cells to scanning notebooks * All notebooks now use nb_utils.py * Remove WIP notebook from commit * Add RGB colour output to sinogram notebooks * Add widget with sliders for Frelon peaksearch data * Add RGB IPF X,Y,Z outputs to HDF, allows viewing with H5Web * py2 compat * Small bugfixes and tidying up * Small changes and bugfixes * Clear output * secret spacegroup number hidden option * secret spacegroup number hidden option * secret spacegroup number hidden option * secret spacegroup number hidden option * Correct very dodgy default cosine tolerance * Correct grain translation sign problems * Correct dodgy do_index scan * Handle missing samples in skips_dict, dset lists are now sorted too * More accurately find user home directory path * Just use default arguments for assemble_label and properties * Force only 2 threads for indexer (confirmed faster) * Improve comments about peak filtration * Correct origin for recon plots * Much faster spatial correction using lut * Lima segmenter now used in interactive notebook * Add howmany to Eiger interactive segmenter * Segment everything in parallel on SLURM * Cleaned-up local PBP indexer * Correct PBP results filename * Fix dodgy pbp name * Fix PBP output name for real this time * fixes #236 * fixes #236 * Correct home directory searching * added options to score_all_pairs for ring selection * added options to score_all_pairs for ring selection * Optionally scale sinogram rows by ring current * Forgot to also add to batch cells * Correct dodgy minor phase grain positioning * nb_utils function for ring current correction per scan * H5py in nb_utils * Make ring current correction optional and separate * More sensible default dstol * tag maintenance release --------- Co-authored-by: Jean-Baptiste Jacob <[email protected]> Co-authored-by: jadball <[email protected]> Co-authored-by: Thomas VINCENT <[email protected]> Co-authored-by: James Ball <[email protected]> Co-authored-by: Jerome Kieffer <[email protected]> Co-authored-by: Jonathan Wright <[email protected]>
- Loading branch information