SymPy Plot Backends v4.0.0

What's Changed

• Refactoring of the plotting module in order to use the param module. While this add a new mandatory dependency (previously, it was installed and used only by the interactive module), it brings many benefits:

  • Clearer code base, in particular inside the spb.series modules, becaue each attributes is now very well documented.
  • Automatic and complete generation of the documentation when the module is imported, on all classes and plotting function. The documentation is shown when the user executes help(plot_function) or plot_function?. All available parameters will be shown.
  • Capability of adding widgets related to attributes of data series.

• Improvements to the function graphics() and the backends:

  • Improved the handling of grid lines thanks to the keyword arguments grid and minor_grid. These parameters can be:

    • boolean: toggle the visibility of major and minor grid lines.
    • dict: keyword arguments used to customize the grid lines.

  • Added capability to set axis ticks to multiple of some quantity (for example pi/2) thanks to the keyword arguments x_ticks_formatter and y_ticks_formatter of the function graphics(). A few preconfigured formatter are available as well, like multiples_of_pi_over_2, multiples_of_pi_over_3, etc.

  • Added hooks keyword argument: it accepts a list of user defined functions that are going to further customize the appearance of the plot. For example, users can change the tick labels on the colorbars, etc.

  • Added app keyword argument: if set to True, an interactive application will be shown, with widgets that allows to customize the numerical data generation process for each data series. Look at the documentation of domain_coloring to see an example.

  • Added use_latex_on_widgets keyword argument, which clearly separates the scope with respect to use_latex.

• Improvements to spb.series (numerical data generation):

  • Added support for plotting summations with infinite bounds on all series (thus, all plotting function as well).

  • Improved evaluation of symbolic expressions producing very large numbers.

  • Improved logic dealing with lambdification in order minimize the time spent in this stage.

  • BREAKING: Removed the adaptive algorithm from line, line_parametric_2d, line_parametric_3d, surface (and their respective counterpars, plot, plot_parametric, plot3d_parametric, plot3d).
    Main motivations were:

    1. easier to use the plotting module, as there are 3/4 keyword arguments less to be worried about.
    2. cleaner and simpler code base.
    3. easier to implement new features (interactive ones), that only works with the uniform evaluation strategy.

  • BREAKING: Split the GeometrySeries into Geometry2DSeries and Geometry3DSeries for better code separation.

  • BREAKING: Changed the init signature of NicholsLineSeries.

  • Added attribute Geometry2DSeries.range_x which allows to plot a Line2D in the specified range along the x-axis.

  • Added attribute LineOver1DRangeSeries.poles_rendering_kw, which is a dictionary of keyword arguments passed to the specific plotting library renderer in order to customize the appearance of vertical lines representing essential discontinuities.

  • Renamed attribute is_point to is_scatter to better represent its intention.

  • Fixed bugs:

    • with the numerical algorithm about singularities detection of 2D parametric lines.
    • with the algorithm used to insert exclusion points through the exclude keyword argument.

• BREAKING: remove tp keyword argument from step_response, ramp_response, impulse_response.