Autolayout Grid

[buhtz]

To my understanding all grid classes derived from Grid are based on data. There purpose is not only to organize and layout multiple plots in a visual grid. Their purpose seems to be taking the relationship between subplot's data into account.

I need a grid with less features and without being coupled to data. I just need the layout feature without the need of importing something from matplotlib.

grid = AutoGrid(columns=2)

cmap_list = ['RdYlGn', 'bwr', 'PuOr', 'BrBG', 'PRGn', 'coolwarm']

for cmap in cmap_list:
    ax = create_sub_axes(mydata, cmap)
    grid.add(ax)

grid.figure.show()

Another variant would be: grid = AutoGrid(rows=2) or grid = AutoGrid(columns=2, rows=2). The latter should raise exceptions if more than columsn * rows (4) axes are added to it. An alternative name would be FlowGrid.

[mwaskom]

From your example I have no idea what you're expecting the result to be but if you just want a grid of axes with no data, you should be using matplotlib directly.

[buhtz]

The result is a grid with 4 subplots in it. What data they show is not important to the grid. That is the point.
It is a Grid just to layout plots totally independent from the data.

And using something else is not an option. Exactly that is the problem.
I want to use seaborn and nothing else. Importing matplotlib, plotly etc are workarounds then. This mess up my code and make it harder to maintain.