Improve Python API

[flekschas]

This is work in progress but I want to document it somewhere:

After discussions with @nvictus I am thinking to change the Python API of plot() to be more intuitive and support dataframes out of the box. The API is inspired by https://seaborn.pydata.org/generated/seaborn.scatterplot.html

jscatter.plot(
  x = str or ndarray,
  y = str or ndarray,
  # Optional:
  data = dataframe, # a dataframe
  color = str or ndarray, # a single color
  color_by = str or ndarray, # color encoding
  color_norm = tuple or matplotlib.colors.Normalize, # normalization strategy for color_by
  color_order = list, # in case of categorical color encoding this specifies the color order
  color_map = list or ndarray, # a list of colors to map color_by to
  size = int or float, # a single point size
  size_by = str or ndarray, # size encoding
  size_norm = tuple or matplotlib.colors.Normalize, # normalization strategy for size_by
  size_order = list, # in case of categorical size encoding this specifies the size order
  size_map = list or ndarray, # a list of size to map size_by to
  connect_by = str or ndarray, # a categorical 
  connect_order = str or ndarray, # order in which the points are to be connected
  opacity = float or str, # a static opacity value that overrides the color's alpha value or 'auto'
)

[flekschas]

Oops, I forgot to close when #4 got merged. The new API is implemented.