Running with spatial patches, hdf5 files

[AmadeusMJY]

Hi suite2p!
I'm recently trying to manually separated a single calcium video into several spatial patches, running them and summing the results. (bec we have a ~300G video, and it seems memory needed are more than twice of the file size, nearly exceeding our server memory ~700G) I wonder which results are needed to be further modified because of the patch processing (like, stat/med evidently) for a reliable summed result?

By the way, we are using hdf5 files. Is there any way to conveniently input spatially separated parts of f['calcium_raw'] (our dataset for video storage) to suite2p? Separate one hdf5 file into several files are somehow clumsy.

Thanks for your time!