Pyxem

===========================

Added
-----
- Added damp_extrapolate_to_zero to ReducedIntensity1D
- Added in deprecation wrapper class to wrap deprecated functions in pyxem.
- Center-of-mass algorithm added to get_direct_beam_position (845)
- Added `VectorSignal1D` class to handle 1 dimensional signals
- Added kwargs to find_beam_offset_cross_correlation allowing for parameters
to be passed to `phase_cross_correlation` (907)
- Added `LazyVirtualDarkField` signal

Changed
-------
- Sklearn is now required to be on the 1.0 series.
- Changed `set_signal_dimension` to `Hyperspy.signals.BaseSignal.transpose`
- Moved code from `VectorSignal` to `VectorSignal2D`. Change is more inline with stated dimensions
- `VectorSignal` pixel_calibration deprecated and replaced with scales.
- Fixed bugs resulting from API change in hyperspy/hyperspy3045. Markers explicitly initialized
- DiffractionVectors.get_diffraction_pixels_map returns a ragged signal
- VirtualDarkFieldImage.get_vdf_segment changed to properly handle setting of axes
- Increased minimal version of scikit-image to >= 0.19.0
- Increased minimal version of Matplotlib to >= 3.3

Fixed
-----
- Fixed a factor of 1/2 missing in ScatteringFitComponentXTables
- Fixed error related to `DiffractionSignal2D.apply_affine_transformation` when multiple affine transformations are given. (870)
- Bugfix related to Numpy 1.24.0. Strict array creation with dtype=object is needed
for ragged arrays. (880 & 881)
- Bug fix for doubling of inplane rotation in template matching. (905 & 853)
- Bug fix for filtering vectors using a basis and DBSCAN
- Bug fix for passing vector attributes when signal is copied or changed.

===========================

Changed
-------
- Increase minimal version of orix to >= 0.9.
- Increase minimal version of diffsims to >= 0.5.

Fixed
-----
- Fix bug in `get_DisplacementGradientMap` (852)
- Fix template matching bugs (originally fixed in 771 but omitted from 0.14 series by accident)

===========================

Added
-----
- Getting and plot integrated intensity now support signals containing nan (722)
- Add Symmetry1D signal class and symmetry analysis methods (724)
- BeamShift class, which includes the `make_linear_plane` method for better correction of the beam shift when scanning large regions in STEM (746)
- Add unit testing of docstring examples (766)
- Add function for optimizing calibration of SPED data (785)
- Add function for creating a orix CrystalMap from indexation results (794)
- Speed optimizations for the fast template matching on CPU and GPU, improving speeds by 200% and 40% respectively (796)
- Added the ability to determine the center and ellipticity using the `determine_ellipse` function.

Removed
-------
- lazy_* virtual imaging has been removed, use get_integrated_intensity (722)
- `big_data_utils` has been removed as this is not the modern way of attacking this problem
- similarly, `TemplateIndexationGenerator` has been removed as the Accelerated approach is far better (823)

Fixed
^^^^^
- Symmetry STEM Class updated to allow for better interpolation and lazy operation. (809)
- Generalized plotting diffraction vectors on ND stacks of images (783)
- Small bugfix with dask/cuda scheduler to prevent running out of VRAM (779)
- Bugfix:AzimuthalIntegral1D accepts masks and uses updated `map` function (826)

Deprecated
^^^^^^^^^^
- The `lazy_result` keyword, which has been changed to `lazy_output` to conform to similar keyword in HyperSpy

Changed
^^^^^^^
- For developers: HyperSpy's `.map` function will now be used to process big datasets, instead of pyXem's `process_dask_array`

===========================

The code contained in this version is identical to 0.14.1, the release was
recreated to fix an error with the Zenodo files.

===========================

Added
-----
- Code now support python 3.9
- Code now runs on hyperspy 1.6.2

Fixed
-----
- np.bool replaced by bool
- np.object replaced by object

===========================

Fixed
-----
- load_mib (734)
- correct_bad_pixels now returns the same result when lazy/not-lazy (bug 723, fix 735)
- mirrored templates now correctly dealt with in radial template matching (740)
- further bugfixes for AcceleratedIndexationGenerator (744)
- a k-space error effecting azimuthal integration (738)
- bug in .to_crystal_map()

Deprecated
----------
- lazy_virtual_bright_field, use get_integrated_intensity instead
- lazy_virtual_dark_field, use get_integrated_intensity instead