Currently, even if we know the current iteration and the total number of iterations an algorithm takes on the server side, we cannot easily share this information with the client. Instead, the default "ininite progress bar" is used until the algorithm finishes running.
To improve this, we could create a Progress data layer which would be used like:
for k in range(max_iter):
...
yield sk.Progress(k, max_iter)In this case, the client would react by updating the progress bar in the UI (or tqdm) at each iteration. We could also use this method to handle progress updates in the tiling case.
Currently, even if we know the current iteration and the total number of iterations an algorithm takes on the server side, we cannot easily share this information with the client. Instead, the default "ininite progress bar" is used until the algorithm finishes running.
To improve this, we could create a
Progressdata layer which would be used like:In this case, the client would react by updating the progress bar in the UI (or tqdm) at each iteration. We could also use this method to handle progress updates in the tiling case.