Bcdi

--------------

* Remove support for Python 3.6 and Python 3.7

* Use exact versions of dependencies in `setup.py` and `requirements.txt` to improve the
reproducibility of results.

* Bug: use explicitely the figure handle to save figures in `bcdi_strain.py` and
`bcdi_preprocessing_BCDI.py`

* Refactor: implement correctly the `__repr__` methods so that eval(repr(instance)) is
itself an instance.

--------------

* Refactor: open files within context managers in `utilities.load_file`.

* Create the class ContextFile implementing the context manager protocol. The method
`loader.create_logfile` returns an instance of ContextFile, allowing its safe usage
in other methods via the decorators `safeload` and `safeload_static`.

* Refactor the signature of `loader.create_logfile`, so that all overriding methods
share the same signature with the overriden base class abstract method.

* Implement the function `utilities.unpack_array`, to avoid hard-coding array slices
in methods using motor positions.

* Feature: add support for the BLISS flavor of ID01 beamline.

--------------

* Refactor: allow the command line parameters to be None. ConfigParser can now be
instantiated providing the path to the config file only.

* Bug: use the binned detector pixel size to calculate the transformation matrices

--------------

* Move the configuration files and example modes data to a subpackage ``bcdi.examples``
so that they are distributed with the builds and wheels.

* Move the run functions from `bcdi_strain.py` and `bcdi_preprocessing_BCDI.py` to new
dedicated modules `postprocessing.postprocessing_runner` and
`preprocessing.preprocessing_runner`, so that it can be imported.

* Add an example in the documentation, using the dataset ID182 from the CXIDB.

* Bug with interpolation with xrayutilities: use the correct number of values to unpack
from the call to `beamline.Beamline.process_position` (the detector distance was
missing)

* Allow the correction of detector angles directly in `bcdi_preprocessing_BCDI.py` and
`bcdi_strain.py`. The user can either provide the Bragg peak position, or this one
will be calculated from the provided direct beam position and the setup geometry.
The user can still provide already corrected detector angles.

* Bug: use the unbinned detector pixel size in `detector.create_roi`,
`diffractometer.init_data_mask` and `diffractometer.load_check_dataset` instead of
the calculated pixel size, which includes the preprocessing binning.

* Rename the parameters for the definition of the detector ROI for data loading
`x_bragg` and `y_bragg` as `center_roi_x` and `center_roi_y`, to avoid the confusion
with the Bragg peak position.

--------------

* Allow lazy loading of the experimental parameters: energy, detector distance, tilt
angle, detector inplane and out-of-plane angles. If not provided by the user, the
values will be looked for in the log/spec file. An exception is raised if they are
not available.

* Implement the chi circle at 34ID-C and update the calculation of the transformation
matrix.

* Add a task in `doit` to check for broken external links in the documentation.

* The methods setup.ortho_directspace and setup.ortho_reciprocal now return also the
transformation matrix from the detector frame to the laboratory frame.

* Support APS 34ID-C for data preprocessing (loading of TIFF images).

* Update the parameter "linearity_function" in preprocessing. Now this can be None (if
unused) or a sequence of 5 real numbers corresponding to the coefficients of a 4th
order polynomial.

* Implement a parser for YAML config files. Now the scripts ``bcdi_strain.py`` and
``bcdi_preprocess_BCDI.py`` can be run like scripts, from the command line, with
optional command line arguments.

-------------

* Bug: apply correctly the non-linearity correction function to the detector frames
(typo in the function name).

* Modify the behavior of the parameter `specfile_name`: for beamlines relying on a
separate file for logging motor positions (specfiles at ID01 and 34ID, fio file at
P10), the user can provide the full path to the file

* Implement loading motor positions from a specfile at 34ID.

* Add mypy for type checking in doit and GitHub workflows.

* Bug: correct the detector horizontal direction in Beamline34ID, it was flipped.

* Rename the parameter data_dirname to data_dir for the function setup.init_paths.
Now the user can provide directly the path to the data directory.

* Move all functions related to dataset alignment in the module
``utils.image_registration.py`` and create generic functions

* Enable preprocessing d2scan using xrayutilities for ID01. The parameter `follow_bragg`
become obsolete and is removed.

* Add the module ``simulation.supportMaker.py``, which allows to create a support using
polygons. Lengths can be defined either in pixels or in nanometers.

* Merge the subpackage facet_recognition into postprocessing and rename the module
``facet_utils.py`` to ``facet_recognition.py``.

* Add the list of publications related to the package in the documentation.

* Add class diagrams in the documentation using sphinxcontrib-mermaid.

* Solve issues with numpy when building the documentation (enable *Use system packages*
in ReadTheDocs advanced settings).