You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At present flatten=False it returns a list of lists, which is not very numerically efficient. Instead is should return as CSR-style, so instead of [ [0, 2, 4], [4, 6]] it returns ([0, 2, 4, 4, 6], [0, 3]) where the second element of the tuple is indices into the first element where each new row starts. You could get each set of neighbors with ind, ptr = <crs-style data> so ind[ptr[0]:ptr[1]]
The text was updated successfully, but these errors were encountered:
This would really hurt readability. Maybe a better idea would be to add an additional option or function to get this CSR style if needed for performance?
At present
flatten=False
it returns a list of lists, which is not very numerically efficient. Instead is should return as CSR-style, so instead of[ [0, 2, 4], [4, 6]]
it returns([0, 2, 4, 4, 6], [0, 3])
where the second element of the tuple is indices into the first element where each new row starts. You could get each set of neighbors withind, ptr = <crs-style data>
soind[ptr[0]:ptr[1]]
The text was updated successfully, but these errors were encountered: