frelon_peaksearch process function split into two functions #395
Conversation
ImageD11/frelon_peaksearch.py
Outdated
| @@ -360,6 +360,95 @@ def guess_bg(ds, scan_number=0, start=0, step=9, n=None): | |||
| PKCOL = [getattr(ImageD11.cImageD11, p) for p in PKSAVE] | |||
|
|
|||
|
|
|||
| def collect_all_frames_peaks( | |||
| dataset, | |||
There was a problem hiding this comment.
Perhaps a bit early but as @jadball suggested, we should not need the Dataset class in the future.
We can give the masterfile, frames_dataset and omega_angles as arguments of the function so that we can reuse it without further changes when we remove the use of the Dataset class.
There was a problem hiding this comment.
correct, this will make future compatible!
| return peaks_2d_dict, num_peaks | ||
|
|
||
|
|
||
| def process_dataset_for_peaks_columnfile( |
There was a problem hiding this comment.
This is essentialy a reimplementation of the process function below right ?
Should we use the new functions directly in the process function instead of creating a new process function ? What do you think @jadball ?
There was a problem hiding this comment.
Maybe I missed some discussion prior to this? The 'importScan.py' code looks like an extraction/refactoring of dataset.py but it is not imported or used in frelon_peaksearch.py. Is this intended to be part of this PR?
Your type hints for CI seem to require https://peps.python.org/pep-0484/#suggested-syntax-for-python-2-7-and-straddling-code
I wrote this Import_scan to do the segmentation with minimal dataset class method but subset of dataset class method, to make the columnfile treatable from its own class, instead of doing it in dataset class I will remote it now, and once I test it completely I will ask for merge in future, sorry of the confusion I made |
|
No problem - if I were you, I would be tempted to run the segmentation starting just from the bliss "master.h5 + scan" and see if you can drop everything else from our dataset class. It was intended for assembling 2D sinograms from many 1D scans... |
Yes. I don't think it was intended for this PR. @abmajith As discussed this week, you don't have to close the PR: keep 633012d only, which the commit that we discussed, in this branch. You could then keep the other commits in another branch in the meantime. |
Hello @loichuder, @jadball
I split the frelon_peak search process function into two, first for running the segment on frames in parallel,
and then merging the peaks found on each frames, so on.