Integrate QC report generation into the workflow #40
Conversation
update to accommodate differing numbers of stains/channels (khanlab#39)
First Merge
(will make sure poetry deps in spimprep-deps ask for ^1.2.0 now)
note: still need to put the css and js in place.. would be nice to have these embedded in the html, maybe that's easier if we use jinja templating?
|
I added the commits I made to fix some things and an example of splitting off the rule.. |
…volume rendering json in the z-direction
|
Code looks good to me, but some documentation could be better. Can you add documentation to briefly describe how to run the .js files? In docs/ or/and using docstring """ """ at the top of the file (as in PEP 257) would be fine. Also may I ask what's the purpose of the dataset created by testing/create_test_dataset.py? I was looking around but not sure what it's used in. I think it may be good to add a description of this as docstrings as well. Other .py files can benefit from docstrings as well but I think these two will be most helpful. Edit: Just saw the issue to add documentation - if it's not part of the PR and will be done in the future then I believe we can merge the PR now. |
|
I can add the documentation for running the js files in this pull request. And the create_test_dataset.py is used to create a subset of a large raw dataset to be able to test the workflow quickly. |
…configurations for the reports
| shutil.move(resource_dir / "volRender.html", html_dest) | ||
|
|
||
| # Get most downsampled ome-zarr image | ||
| ds_z = ZarrNii.from_path(ome_data,level=5, channels=[0,1]) |
There was a problem hiding this comment.
hard-coding the channels here fails when we have a different number of channels, but I think the best solution is to make a fix in zarrnii so that the default channels selects all of them (instead of just channel 0).. I'll make a note there to fix it, then we can change this line in a future PR..
needed because of basicpy update
|
@BenGros this looks good to merge now, when you're ready you can do a Squash and Merge to clean the history, and provide a commit message and summarizes the PR. |
QC reports are generated for flat-field correction, whole slice viewing and a 3D volume rendered model. These reports are accessible through a locally run HTML file. Also added a script for producing test datasets that are subsets of an original dataset.
The ff correction and whole slice viewing use locally stored images that are created.
The 3D viewer generates a JSON file containing the data and generates the model from that.