I had been still getting the AssertionError: Arrays differ by non-integer cycles problem when trying to tile. I tracked this when I wrote out each spatial unwrapping result to a numpy array within the future, it perfectly matched the spurt CLI. But then the HDF5 saved data was different.
The cause: as_completed returns data out of order (possibly). This leads to random row shuffling of uw_data.
Still not quite sure how doing the single worker didn't fix this when we checked earlier.
(Also: I missed adding separate options to the CLI in the last PR)
I had been still getting the
AssertionError: Arrays differ by non-integer cyclesproblem when trying to tile. I tracked this when I wrote out each spatial unwrapping result to a numpy array within the future, it perfectly matched the spurt CLI. But then the HDF5 saved data was different.The cause:
as_completedreturns data out of order (possibly). This leads to random row shuffling ofuw_data.Still not quite sure how doing the single worker didn't fix this when we checked earlier.
(Also: I missed adding separate options to the CLI in the last PR)