Nexpy

Enhancements
- Addition of Gaussian smoothing, using the astropy convolution function. This is available through the 'convolve' option in the Signal tab. The standard deviation of the Gaussian can be changed typing `plotview.smooth=<n>` in the shell, where `n` is the number of pixels (2 by default).
- Support for symmetric color maps, in which the minimum signal value is fixed to the negative of the maximum. These are invoked by choosing one of the divergent color maps, which are now separated at the bottom of the color map menu in the Signal tab. The `linthresh` and `linscale` parameters can be set by calling `plotview.symlog(<linthresh>, <linscale>, <vmax>)` in the shell. Calling `symlog` without any parameters resets the defaults (`linthresh=vmax/10`, `linscale=0.1`).
- Z-axis values now set to values closest to 0 when selecting the default slice for a 3- or higher-dimensional plot, _i.e._, with symmetric axes, the z-axis values will default to the middle. In our experience, this is nearly always preferable to the old default of plotting the zero-index slice.
- New 'Set Default' menu item, which allows default data and/or entry groups to be selected for subsequent plotting. This uses the 'default' attribute on the NXdata or NXentry groups, as defined by the NeXus standard.
- Configuration of grid colors and lifestyles in the Customize Panel.
- Addition of a new 'Import Image' option, for importing image files that contain RGBA data, _e.g._, JPEG or PNG files.
- Allow addition of attributes to groups in the 'Add Data' dialog.

Bug Fixes
- Fixes a bug in setting the limits with infinities in the data.
- Fixes bugs with changing and skewing axes.
- Fixes problems caused by the failure of Matplotlib to update the colorbar ticks and label formatting when the 2D image normalization is changed, _e.g._, to log scaling.
- Fixes a bug with the 'Change Plot Limits' menu item.

- Adds a new tabular view of groups and fields, using the new `View Data` menu. The associated metadata, including attributes, are displayed along with all the data values in 10 x 10 blocks.
- Adds menus to backup and restore NeXus files. Whenever a file is unlocked, the user can choose to store a backup version, which may be restored in the user's NeXpy directory (`~/.nexpy/backups`) if changes to the unlocked version need to be reversed. These backups are stored for three days, and may be read in again or deleted using the 'Manage Backups' menu.
- Various enhancements of the Fit Dialog, allowing both trial parameter estimation and least-squares fitting to be restricted to a range of x-values within the data.
- The scratch workspace, `w0`, is now saved to the user's NeXpy directory (`~/.nexpy`) and can be used to save snippets of NeXus data for reuse in a subsequent session. New menu items will open, purge, and close this scratch file.
- Adds a new command-line option to display the current image using symmetric logarithms, `plotview.symlog`. A number of diverging color maps have been added to the default list.
- New color maps can be installed from the shell if they are available in the Matplotlib installation.
- Externally linked items can now be renamed.
- All plot windows can be resized to be the same as the main plot window to facilitate comparisons.
- Fixes the behavior of the Projection Panel with reversed axes.
- Uses the new v0.4.4 release of the `nexusformat` package, which fixes a number of issues with external links.

- Data projections are now averaged over the number of summed bins, by default. The original behavior, in which the data are summed, can be restored by checkboxes in the Projection Tab and Projection Panel.
- Fixes a bug causing a crash on launch on some systems because of conflicting PyQt libraries. The environment variable, QT_API, is set when the PyQt library is first imported, to prevent other packages from attempting to import a different library.
- Requires v0.2.2 of the nexusformat package, which fixes a Python 3 bug preventing NXfield and NXgroup divisions.

- This is the first version compatible with both Python 2 and Python 3. It has been tested on v2.7, 3.3, 3.4, and 3.5. Python 2.7 and 3.3 are the minimum versions for Python 2 and 3, respectively.
- Restores default tick mark locations. When implementing skewed axes, the default tick mark locations were inadvertently restricted to integer values. This restores the Matplotlib defaults. A new command-line function, plotview.locator, allows these to be modified.
- Adds the ability to check the NeXpy version from the command line without launching the GUI by typing `nexpy --version`.
- Updates requirement for nexusformat to v0.2.1 to improve Python 3 compatibility.

* Provides an option to store lock files in a directory designated by an environment variable or preference parameter.
* Creates `nxgetconfig` and `nxsetconfig` functions for handling parameters, such as default compression and memory limits. The older `get` and `set` functions are kept for backward compatibility.

* Adds a `transpose` function to NXdata. The syntax is like the NumPy and NXfield `transpose` functions, but the function also permutes the axes.
* The `nxsignal` property returns a NXfield if it is the sole member of a NXdata group, even if the `signal` attribute is not set.
* Adds `moment`, `mean`, `var`, and `std` functions to NXfields, using the array indices as the x-axis values.
* Fixes compatibility with Matplotlib v3.6.