Nexpy

- Adds the ability to plot in 2D with skewed axes, _i.e._, where the x and y axes are not orthogonal. This is important when reciprocal space data are stored in a mesh that is diagonal in (_h,k,l_), but the unit cell has non-orthogonal axes.
- A new "Customize Panel" replaces the default Matplotlib edit dialog, with fields tailored to NeXpy. For 2D plots, the skew angle, aspect ratio, and grid options can be set in the panel. For 1D plots, marker and line styles can be modified.
- Redefines the default color map to use the forthcoming Matplotlib default, 'viridis'. This is one of four new perceptually linear color maps that render intensity variations with less distortion.
- Right-clicking on a plot restores the axis limits without resetting the signal limits. Signal limits can be reset by clicking on the Home button in the options tab.
- Fixes an incompatibility with the new Jupyter Console, which prevented the shell from editing multiline code.
- Fixes an incompatibility with Matplotlib v1.5.0, which prevented rectangular zooms from working.
- Fixes an incompatibility with lmfit-py v0.9.0, which prevented 1D fits from working.
- Fixes a Windows bug where the plotting tabs had zero height.
- Handles missing imports more sensibly.
- Adds the .cxi extension to the list of recognized files in the default Open dialog.

- A new 'Open Recent' menu is added to the 'File' menu to access the 20 most recently accessed NeXus files.
- Fixes a bug in replotting images with irregular bin sizes. Unfortunately, this has required reverting to using the Matplotlib pcolormesh module, which does not have an interpolation option, for plotting irregular grids. The requested interpolation method is cached and restored when a regular grid is plotted again.
- Sets the PyQt4 version to v2 to prevent problems using PyQt4-based IDEs to debug NeXpy.

- The previous release did not explicitly require IPython v4.0.0, but this is necessary to run with the new Jupyter console, at least in some Python configurations. This is the only change.

- With the release of IPython 4.0.0, the Qt console code has been transferred to the language-agnostic [Jupyter project](http://jupyter.org) along with an extensive refactoring of the code. NeXpy has now been updated to use the new jupyter package, which is listed as a new dependency.
- A dropdown menu to apply any of the available matplotlib interpolation methods to the current plot has been added. If the grid is not equally spaced, then the only method is 'bilinear', but there are many more options for plots using 'imshow'.
- The display of line numbers has been removed from the script editor except when a Mac is being used. This is because there is a discrepancy in the line heights for the line number and text editor boxes in other PyQt implementations. Line numbers will be restored when this has been fixed.

- Scripts are now opened as tabs in a single script editor window and displayed with line numbers.
- Fixes a bug when switching axes in a three- or more-dimensional data set plotted using pcolormesh.

- Bug fixes to improve PyQt4/PySide compatibility. Keyword arguments, allowed by PySide, were replaced by the positional arguments required by PyQt4.
- Improvement to the plugin utilities. When using GridParameters, all parameters can now be restored after an unsuccessful refinement, not just those that are varied. Some parameters might be changed by a fit function even when they are not varying, _e.g._, when two parameters are constrained to be equal.