Summary

The current interface is neither very intuitive nor practical nor in line with the traditional gridspec logic.

Proposed fix

Half a note to self, because this is not completely thought through:

• The parameterngrids seems not to be working (usage raises)
• The attributesngrids is not helpful. It's misnomed (naxes would be better), and it's equivalent tolen(grid) orlen(grid.axes_all).
• Not sure what we needgrid.axes_llc for. That seems quite specific.
• The 2D grid arrangementaxes_row,axes_column are lists of lists. In contrastsubplots returns an array. Thus, you need double indexinggrid.axes_row[row][col} and cannot use the multi-indexaxs[row, col].
• Having thedirection parameter andgrid.axes_row,grid.axes_column is basically two redundant approaches towards ordering.

Possible improvements:

• deprecatengrids attribute - replaced byFix ngrids support in axes_grid.Grid(). #27972
• add an attributegrid.axes or maybe evengrid.axs (in analogy tofig, axs = plt.subplots(2, 3)), which is a ndarray of the Axes. Hence one could dogrid.axs[row, column].
• maybe (??) soft-deprecate axes_all, axes_row, axes_col, direction?
• inGrid.__init__ make everything afternrows_ncols kw-only
• would be nice to have a defaultrect, but that requires reorganizing parameter order because that's beforenrows_ncols.
  Might be possible migrate

  def __init__(self, fig, rect, nrows_ncols, *, ...)

  to

  def __init__(self, nrows_ncols, *, rect, ...)

  with reasonable deprecation warnings and catching inappropriate use. - But would have to think this through