What can be improved?
Currently, rolling sums only take a single value for controlling the window.
For a fine-grained control, we need to specify the size of the window summed over per entity. Then, sum_back(x, over=d, within=n) takes a number or an integer parameter, so a window may be a different length. The value likely stays static along dimension d ie. in pypsa terms we don't have different windows sizes per component and snaphot, only per component.
Version
v0.0.0-alpha.13
What can be improved?
Currently, rolling sums only take a single value for controlling the window.
For a fine-grained control, we need to specify the size of the window summed over per entity. Then,
sum_back(x, over=d, within=n)takes a number or an integer parameter, so a window may be a different length. The value likely stays static along dimensiondie. in pypsa terms we don't have different windows sizes percomponentandsnaphot, only percomponent.Version
v0.0.0-alpha.13