chilipp/psyplot: v1.1.0: New xarray Dataarray and Dataset accessors

This new release mainly adds new xarray accossors (psy) for DataArrays and Datasets. Additionally we provide methods to calculate the spatially weighted mean, such as fldmean, fldstd and fldpctl. Added features are: - The yaxis_inverted and xaxis_inverted is now considered when loading and saving a matplotlib axes - Added the seaborn-style command line argument - Added the concat_dim command line argument - Added the plot attribute to the DataArray and Dataset accessors. It is now possible to plot directly from the dataset and the data array - Added requires_replot attribute for the Formatoption class. If this attribute is True and the formatoption is contained in an update, it is the same as calling Plotter.update(replot=True)). - We added support for multifile datasets when saving a project. Multifile datasets are datasets that have been opened with, e.g. psyplot.data.open_mfdataset or psyplot.project.plot..<plotmethod>(..., mfmode=True). This however does not always work with datasets opened with xarray.open_mfdataset. In these cases, you have to set the Dataset.psy._concat_dim attribute manually - Added the chname parameter when loading a project. This parameter can be used to display another variable from the dataset than the one stored in the psyplot project file Added the gridweights, fldmean, fldstd and fldpctl methods to the psy DataArray accessor to calculate weighted means, standard deviations and percentiles over the spatial dimensions (x- and y). - Added the additional_children and additional_dependencies parameters to the Formatoption intialization. These parameters can be used to provide additional children for a formatoption for one plotter class - We added the psyplot.plotter.Formatoption.get_fmt_widget method which can be implemented to insert widgets in the formatoptions widget of the graphical user interfac