To get the most out of brainrender, your data should be registered to one of the atlases supported by brainglobe's AtlasAPI. If you've used brainglobe's software for processing your raw data (e.g. brainreg) your data will already be registered and you need not worry about any of this. If not, then some steps are necessary for registering your data.
Brainglobe's AtlasAPI relies on bg-space for transforming data (e.g. image stacks) so that they are all oriented the same way. Bg-space provides a convenient naming convection to define the orientation of your data based on where the origin is and the direction that the three main axes (first three dimensions of your image data) point towards.
The process of transforming data from one axes system requires knowing the "space" of your target (i.e. of brainrender's atlas data) and of the source (your data). The orientation of your data depends on your experimental set up and subsequent pre-processing steps. To know what brainrender's target space is:
from brainrender import Scenescene = Scene()print(scene.atlas.space)"""<BGSpace AnatomicalSpace object>origin: ('Anterior', 'Superior', 'Right')sections: ('Frontal plane', 'Horizontal plane', 'Sagittal plane')shape: (528, 320, 456)"""
Check bg-space documentation for more details: https://github.com/brainglobe/bg-space.
The section above described how to sort your image axes to that the coordinates order matches brainrender's. However, you might need additional steps to ensure that your data are registered to the atlases: your data might be at a different resolution or to a different offset. Resolution refers to how many microns (all units in brainrender as in microns) the side of the voxels in your image correspond to. Offset referes to the fact that the origin of your image might be offset from the origin of the atlas space (e.g. if you didn't image the entire brain).
Here too brainglobe's bg-space provides tools to address mismatches in these two aspects.