Bug #4204
closedcatch error in validation and commit and move on
0%
Description
ctf estimation is used in live data collection and it makes stopping appionLoop when validation has problem a bad place to stop.
Here are examples of them which I will use a try-except to bypass.
16may23b_00131fcn.mrc
Traceback (most recent call last):
File "/home/acheng/myami/appion/bin/ctffind4.py", line 353, in <module>
imgLoop.run()
File "/home/acheng/myami/appion/appionlib/appionLoop2.py", line 95, in run
self.loopCommitToDatabase(imgdata)
File "/home/acheng/myami/appion/appionlib/appionLoop2.py", line 153, in loopCommitToDatabase
return self.commitToDatabase(imgdata)
File "/home/acheng/myami/appion/bin/ctffind4.py", line 306, in commitToDatabase
ctfinsert.validateAndInsertCTFData(imgdata, self.ctfvalues, self.ctfrun, self.params['rundir'])
File "/home/acheng/myami/appion/appionlib/apCtf/ctfinsert.py", line 45, in validateAndInsertCTFData
ctfvalues = runCTFdisplayTools(imgdata, ctfvalues, opimagedir, fftpath, fftfreq)
File "/home/acheng/myami/appion/appionlib/apCtf/ctfinsert.py", line 80, in runCTFdisplayTools
ctfdisplaydict = ctfdisplay.makeCtfImages(imgdata, ctfvalues, fftpath, fftfreq)
File "/home/acheng/myami/appion/appionlib/apCtf/ctfdisplay.py", line 1436, in makeCtfImages
ctfdisplaydict = a.CTFpowerspec(imgdata, ctfdata, fftpath, fftfreq, twod=twod)
File "/home/acheng/myami/appion/appionlib/apCtf/ctfdisplay.py", line 1318, in CTFpowerspec
normpowerspec = self.normalizeCtf(powerspec, twod=twod)
File "/home/acheng/myami/appion/appionlib/apCtf/ctfdisplay.py", line 588, in normalizeCtf
ind5peak2 = numpy.searchsorted(raddata, peakradii[5])
IndexError: index out of bounds
Updated by Neil Voss over 8 years ago
can you upload the image and provide: apix, Cs, and kv
could be another close to focus error; can I get the estimated defocus.
Updated by Neil Voss over 8 years ago
I have a try/except fix, but I have other changes I am actively testing something else...
Updated by Anchi Cheng over 8 years ago
Neil,
The image is a highly astigmatic focus image for coma-free alignment. One axis has zero defocus. ctffind just gave garbage fit. The validation program is rightfully failed. It just need to handle the error so that it can move on.
I am also feeding the program coma-free alignment Zemblin tableau which is not a real image, but there is no way to distinguish the tableau with the real images at this point unless I hide them.
Do you know what is the consequence of me just skipping database insertion ? Could this break later query ? In additional to pass the exception, I wonder if we should start a log of failed images.
Updated by Neil Voss over 8 years ago
I submitted a fix that fixes your specific error. Other errors could cause a problem, but I do not like to put a hard cutoff on the minimum defocus.
Would love a log of failed images that I could test, see #3950