adding functions for 2p alignment #7
Conversation
fimpy/pipeline/alignment.py
Outdated
| @@ -8,6 +8,8 @@ | |||
| shift_stack, | |||
| sobel_stack, | |||
| ) | |||
| from skimage.feature import register_translation | |||
There was a problem hiding this comment.
I thought we switched to https://scikit-image.org/docs/stable/api/skimage.registration.html?highlight=phase_cross_correlation#phase-cross-correlation with the new scikit-image
There was a problem hiding this comment.
True, I missed this. I switched the functions.
| shape_block=(dataset.shape_block[0], 1) + dataset.shape_block[2:], | ||
| ) | ||
|
|
||
| if verbose: |
There was a problem hiding this comment.
we should replace all of those verbose with the logging module (can be done in a later PR)
| """ | ||
|
|
||
| # prepare the destination | ||
| new_dataset = EmptySplitDataset( |
There was a problem hiding this comment.
if resolution is in the original dataset, it would be good to pass it here
There was a problem hiding this comment.
Should I add an optional input variable for the resolution (with None as default)?
There was a problem hiding this comment.
no, it doesn't make sense to insert it here, it should only be passed if the original dataset has it
There was a problem hiding this comment.
Ok, so this request is not clear to me. What does if the original dataset has it means? I thought because fimpy is public it should be general.
There was a problem hiding this comment.
So at some point it was decided that the SplitDataset class optionally contains resolution and dimension order attributes
e.g.
https://github.com/portugueslab/split_dataset/blob/1faa37954a1fe7bcd23021e2f93238f7b94affb5/split_dataset/split_dataset.py#L113
it makes sense that all operations that work on SplitDataset preserve this information. (so if you e.g. align a dataset which has these fields, the aligned dataset will have these attributes as well). A more elegant, consistent solution for all of this would use xarray, but @vigji is still ambivalent over file formats supported by xarray.
There was a problem hiding this comment.
I am ambivalent indeed :D so so far we are just storing the resolution, and if one day we move to xarray supporting formats we have the information to move there the stack easily
There was a problem hiding this comment.
it's not just resolution, dimension order is important too
There was a problem hiding this comment.
Can I merge the branches and deal with this later? I would like to start running alignment of new data
Added the old 2p alignment function to fimpy: