fetching difumo 1024 statistical map can be memory heavy if it is caclculated on the fly

[xgui3783]

as of a000337

sample code to reproduce:

import siibra
_map = siibra.get_map("difumo 1024", "mni152", "statistical")
_map.fetch("Inferior frontal sulcus middle right")

expected result: returns nii

actual result: process Killed.

[xgui3783]

update: it seems memory related.

the code ran successfully after a machine restart, but htop is showing a ram usage of ~10GB.

This is an issue for siibra-api, since it runs siibra-python on a resource constrained pod (~2GB).

[AhmetNSimsek]

Needs to be checked again after #334 is completed.

[AhmetNSimsek]

as of 0588aca, the issue is resolved for local machines. https://github.com/FZJ-INM1-BDA/siibra-python/tree/wip_clarifySparsemap aims to solve this by committing the precomputed SparseIndex to memory instead of writing to the local cache.