-
Notifications
You must be signed in to change notification settings - Fork 22
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Invalid terminology information in tooltip and Slicer crash using segment descriptor file context #135
Comments
@fedorov Any thoughts on this? It's essential for the paper... |
I tested on nightly build 01/06/2017 with using the master dcmqi and master QuantitativeReporting. This worked for me which brings me to the conclusion, that it must be the integration of rapidjson |
I've just got to this now. Quite likely regression due to rapidjson. I'm investigating. |
I need step-by-step instructions to reproduce the problem (simply browsing/changing terminologies in Segment Editor works fine) |
|
Dataset that I used is QIN-HEADNECK-01-0139 from http://slicer.kitware.com/midas3/folder/3771 |
= in the DICOM browser select a series with modality SR from that study that Christian mentioned above |
ok, thanks! |
I also just got the notification. I'll look at the issue with the data you linked. |
Thanks guys! |
Although the email header says 6PM yesterday, I'm pretty sure I just got it. Not sure why this is. |
Sorry for the delay, I needed to rebuild parts of Slicer after updating. I also updated and built DCMQI and QuantitativeReporting. When I load the SR series, I get this error Any ideas what to do? |
Csaba, we introduced a new dependency, which would be resolved if you installed via ExtensionManager. You will need to build and add to the paths this extension: https://github.com/QIICR/Slicer-PETDICOMExtension |
Good to know, thanks! Is it intentional that the name of this extension has "Slicer-" prefix and "Extension" postfix, while the others have neither? (QuantitativeReporting, LongitudinalPETCT, etc) |
After downloading and building the other extension I got past the previous error.
|
I've installed PETDICOMExtension and the issue is the same. |
@cpinter That probably means, that the Segmentation is not properly loaded into Slicer. Are you using the most recent version of QuantitativeReporting and dcmqi? |
Yes |
Other than that, there should be any error output in the log... or the segmentation itself is missing in the SlicerDICOMDatabase |
@che85 could this be related to QIICR/dcmqi#173? I have not tested the latest one, but it worked for me in earlier nightlies. |
@cpinter that extension was developed by the Iowa team, and they follow their own naming conventions. I agree, I don't see a need for those. Maybe @chribaue can comment on this. |
@fedorov no that cannot be related since QIICR/dcmqi#173 is only about the webapps |
If it is the same one that is referenced from SEG ... I will download the nightly and test from scratch now. Building is a slower option, since my slicer build is quite old. |
@che85 sorry - I meant this one QIICR/dcmqi#170. Maybe in the process of changing the attribute names something was broken? |
@fedorov No I am pretty sure that it doesn't have to do with that since I tested the new dcmqi and Quantitative Reporting with an older nightly of Slicer from 01/06/2017 before integration of rapidjson |
@cpinter @fedorov @lassoan I am getting the same error with the latest nightly. I will look into the error messages
|
I thought this used to work ... let me think what I can turn off/change ... |
Strange. Unfortunately the CircleCI check also doesn't work any more (the last 15 or so checks failed the same way although all of them except my commit yesterday should have passed). Maybe we can ask Steve or someone to try to build it as well. I understand it's really urgent to make this work, and Christian only has Mac. FYI I committed another small fix, it should be good now. There is only one very minor issue, namely that when entering QuantitativeReporting after loading the SR I get this warning many times: |
I have no idea what is going on with my build. I checked the manual, and there are no special provisions re SimpleITK on Mac. @che85 can you run a build on your system? Maybe it is because my OS is couple of generations old. Although, I know my build was working recently. |
I got similar ITK-related errors about but when building CTK. I also skipped OS updates, and I stopped being able to build Slicer a few months ago. I'm updating now to whatever animal or national park is the latest, then try from scratch. I won't be able to finish today for sure though. |
I will update, but not before Feb 1 - too much risk of breaking things I will need for the paper! |
Ah, I think I might have found it! I think I actually did have this issue before. I will be trying this suggestion from Brad that worked last time:
Maybe not too late for you! |
It is :( However a build error should be quite easy to fix once you can reproduce. Although I hope the error is gone, as they seemed to be related to two warnings which I fixed since then. |
@cpinter I cannot recall if I have ever seen this error message. |
@cpinter to make you feel better - that trick did not help! |
@che85 Thanks! I'm trying to fix it, but nevertheless, if the Mac build is OK then it will work now. |
It would be useful to limit SimpleITK dependency, as it's a huge component and the inability to debug makes troubleshooting almost impossible on Windows. Of course SimpleITK can be still used, but try to implement modules so that only those features are not available that actually require SimpleITK. |
@pieper If you can easily update and build Slicer on your Mac would you please try that? Thank you! |
I am also going to build Slicer from scratch now on my Mac. |
I was able to build Slicer without problems from Slicer/Slicer@707ee80 |
I've started a fresh mac build of the current master.
…On Fri, Jan 27, 2017 at 3:15 PM, Christian Herz ***@***.***> wrote:
I was able to build Slicer without problems from ***@***.***
<Slicer/Slicer@707ee80>
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#135 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAHsfbJ7WnuF4PcwJMExwIq5PNuBslQ1ks5rWlBfgaJpZM4Ls5UQ>
.
|
My fresh (debug) build seems to work fine. Is there something else to test?
[image: Inline image 1]
…On Fri, Jan 27, 2017 at 6:58 PM, Steve Pieper ***@***.***> wrote:
I've started a fresh mac build of the current master.
On Fri, Jan 27, 2017 at 3:15 PM, Christian Herz ***@***.***>
wrote:
> I was able to build Slicer without problems from ***@***.***
> <Slicer/Slicer@707ee80>
>
> —
> You are receiving this because you were mentioned.
> Reply to this email directly, view it on GitHub
> <#135 (comment)>,
> or mute the thread
> <https://github.com/notifications/unsubscribe-auth/AAHsfbJ7WnuF4PcwJMExwIq5PNuBslQ1ks5rWlBfgaJpZM4Ls5UQ>
> .
>
|
That's great news, thanks! @che85 could also build, so he can test the rest with the proper dataset and extensions. |
@cpinter no I cannot test it with my Slicer build since Slicer is not starting properly on my Mac OS Sierra. Will have to wait for the next nightly build. |
@che85 Do you know why it doesn't start? This is not what I'd call built without problems. |
@cpinter Slicer built 100% without errors. My startup problems already existed months ago... |
Seems like the nightly failed on the build system for Mac: http://slicer.cdash.org/viewBuildError.php?buildid=962482 Interesting that neither @pieper nor my build failed |
It seems to be a ConstMemberIterator / MemberIterator mismatch. Solution is probably to replace RemoveMember by EraseMember as suggested here: Tencent/rapidjson#721 @cpinter Can you give this potential solution a try? (if you cannot test it yourself then just change it and wait for the next nightly) |
I replaced all RemoveMember calls to EraseMember, and it seems to work OK. |
I built Slicer with SimpleITK disabled, in Release mode. Loading works fine, and I see segmentations overlayed. I cannot access QuantitativeReporting UI, since it has a dependency on SimpleITK, but when I try to switch to SegmentEditor, I get a crash. I will now build in Debug to see if I can get any more info on the crash. |
Ok, I am super confused now - now it is working! It was crashing repeatedly, then I discovered it was actually build in debug mode (I guess the flag got reset as I was rebuilding), so I ran it in the debugger, it didn't crash, and now I cannot reproduce the problem. Hope I was just messing something up earlier - fingers crossed for tomorrow nightly! |
Thanks for the news, Andrey! We're actually late from a party now, so I'll then stop trying to reproduce and will get back to it in the evening. I hope it's not broken! |
Csaba, definitely - it makes a lot of sense to wait until tomorrow. Thanks for your efforts! |
Mac build was successful on the factory, so that's good. |
Confirmed, looks good - thanks!! |
Terminology seems not to work properly anymore when LoadTerminologyFromSegmentDescriptorFile/LoadAnatomicContextFromSegmentDescriptorFile are used.
Hovering over a segment within the integrated Segment Editor displays "invalid terminology information"
Also, when creating new measurements and selecting the context which has been loaded from the segment descriptor file, Slicer crashes. [1]
This is not related to the changes I made in #132
My assumption is, that this is related to the changes made using rapidjson instead of jsoncpp.
@cpinter , @lassoan Any ideas?
[1] https://gist.github.com/che85/4d0a772ac5adc03958333c2a16f8f3a1
The text was updated successfully, but these errors were encountered: