Final frame fix #82
Final frame fix #82
Conversation
core/lls_core/models/lattice_data.py
Outdated
| Check final frame, if acquisition is stopped halfway through it causes failures | ||
| This validator will remove a bad final frame | ||
| """ | ||
| final_frame = v[-1] |
There was a problem hiding this comment.
Note that v[-1] will just take the last index along the first axis, but not the other axes. Is this what you want here?
>>> arr
<xarray.DataArray (T: 2, C: 3, Z: 4, Y: 5, X: 6)>
>>> arr[-1]
<xarray.DataArray (C: 3, Z: 4, Y: 5, X: 6)>There was a problem hiding this comment.
Yep that's correct. I only want to check the final frame, that's the only one that should fail in this way, I could probably drop to a single channel as well to save processing time but it's a pretty marginal improvement and this is clearer I think
There was a problem hiding this comment.
In that case, we could reduce the memory usage and possibly clarify what is happening by doing:
| final_frame = v[-1] | |
| final_frame = v.isel(T=-1, C=-1) |
There was a problem hiding this comment.
Yeah that is a bit more explicit, will make that change
There was a problem hiding this comment.
You can click the "commit suggestion" button to use the change directly if you want.
| @@ -76,6 +77,21 @@ def read_image(cls, values: dict): | |||
| # Use the Deskew version of this validator, to do the actual image loading | |||
| return super().read_image(values) | |||
|
|
|||
| @validator("input_image", pre=True, always=True) | |||
There was a problem hiding this comment.
| @validator("input_image", pre=True, always=True) | |
| @validator("input_image", pre=False, always=True) |
| @@ -191,7 +191,7 @@ def read_image(cls, values: dict): | |||
|
|
|||
| # If the image was convertible to AICSImage, we should use the metadata from there | |||
| if aics: | |||
| values["input_image"] = aics.xarray_dask_data | |||
| values["input_image"] = aics.xarray_dask_data | |||
There was a problem hiding this comment.
| values["input_image"] = aics.xarray_dask_data | |
| values["input_image"] = aics.xarray_dask_data |
Dealing with issue #80
This branch also has the fixes for the other issues but all should be able to be merged once approved