Closed jadie1 closed 3 years ago
I think we can drop all the tiny-tests, no? GH Actions runs them on all platforms.
I think we can drop all the tiny-tests, no? GH Actions runs them on all platforms.
Agreed
What are the OSX and Windows issues? We lost the installers?
What are the OSX and Windows issues? We lost the installers?
OSX doesn't have an installer (.app in the comment above), but Windows had one. You just have to click on Downloads to see it.
@cchriste @akenmorris @jadie1 @iyerkrithika21 @archanasri @riddhishb @HeavenlyBerserker @jmozing2 @RichLisonbee
Where are we on the release testing?
@sheryjoe I think we should assign specific testing tasks to each person. Otherwise, we'll witness diffusion of responsibility in play here.
@HeavenlyBerserker As we did in the 6.0 release (#1073), the table in this issue should help each one to "proactively" start testing depending on the system(s) he/she has access to.
I'll test all the use cases on Ubuntu 20.04.2 LTS today.
Diffusion of responsibility is a sociopsychological phenomenon whereby a person is less likely to take responsibility for action or inaction when other bystanders or witnesses are present. Considered a form of attribution, the individual assumes that others either are responsible for taking action or have already done so.
I'll test all the use cases on Ubuntu 20.04.2 LTS today.
Can you edit the table to mark in-progress testing on Ubuntu to avoid duplicate efforts on a similar platform?
@jmozing2 (the only ortho user with Mac) won't be able to test the release today. @MacUsers (@riddhishb @archanasri @cchriste @akenmorris) could you proceed with testing the release on Mac? Please update the table accordingly to avoid duplicate testing on a similar platform. Thanks.
I don't think we have a build yet. #1283 needs to be merged first. Otherwise none of the python stuff works at all, my understanding.
@jadie1 could you test use cases and studio on OpenSUSE?
@HeavenlyBerserker please test studio on Ubuntu.
@iyerkrithika21 I recall that you have a window machine, could you take the lead on windows testing?
I don't think we have a build yet. #1283 needs to be merged first. Otherwise none of the python stuff works at all, my understanding.
@akenmorris Should this be tackled first? When will this be ready? Should I test Studio in the meantime, or should I wait?
@cchriste what is the status of #1283?
I don't think we have a build yet. #1283 needs to be merged first. Otherwise none of the python stuff works at all, my understanding.
@akenmorris Should this be tackled first? When will this be ready? Should I test Studio in the meantime, or should I wait?
@HeavenlyBerserker I think you should be able to start testing Studio.
@sheryjoe , yes you can test Studio now. The artifacts from PR #1283 should work for Python. I'd like to get it merged and create an RC.
@akenmorris Should I test master? There is no release_v6.1 branch.
@HeavenlyBerserker , sure.
How do we tell correctness?
Anything else we need to test for?
Also, we should include instructions for what to look for in tests in the future. Otherwise, I can only make sure it runs and the results are somewhat reasonable.
Edit: I added instructions to the first comment. Please let me know if we have to look for anything else.
Prove me wrong.
Correctness should be typically objective and measurable, especially wrt unit testing. But we currently have many tests deemed successful if they simply finish without crashing.
Everyone: please file specific issues for things like this as you test.
On Mon, Jun 14, 2021 at 12:45 PM Hong Xu @.***> wrote:
How do we tell correctness?
- I'm assuming that for Studio, we run grooming and optimize with default parameters and make sure the particle sampling, correspondence, and animate are reasonable?
- For use cases, we just make sure sampling, correspondence, and animate are reasonable, right? Anything else we need to test for?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/SCIInstitute/ShapeWorks/issues/1273#issuecomment-860910018, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAJT3ELIRGQ4HR6ST32BXPDTSZE4FANCNFSM46KWIFZQ .
1280 (pr #1283) fixed and merged to master: conda_installs now installs the python module on all platforms.
Prove me wrong.
Kudos to @akenmorris: #1297 I have a fix demonstrated locally, pull request coming soon. Clean installation testing for all platforms will need to be re-tested, along with at least the ability to execute RunUseCase for all platforms (the results will not be different; just want to make sure they can still import the python module).
The new ShapeWorks installation is now ready to be tested (use rc3 tag). For the Python module It's much simpler and no longer requires duplication of libraries or clever modification of runpaths. A new parameter was added to RunUseCase.py, --shapeworks_source
, that developers can use to ensure their module modifications are imported (they should also be sure to use --shapeworks_path
to indicate the path to their build/bin directory).
Update from Tanner who is on windows and unable to edit the issue: Tested feature map and mesh grooming in studio and got the results in the attached pictures. I'm not sure if the feature map picture is just a visual bug (or if that's how its supposed to look), but in the mesh grooming test the particles didn't line up with the model. This occurred after changing some grooming parameters and rerunning it.
Thanks Tanner! I've created an issue for this: #1309
Updated links for RC3.
Update links for RC5.
@cchriste I followed the instructions for the Linux binary build for RC8 and it worked (use cases and notebooks ran succesfully). When I did the same thing for RC10 it installed successfully but when I ran a use case I got this error:
I do have the environment activated, not sure how to debug further.
@jadie1 , can you try RC12?
https://github.com/SCIInstitute/ShapeWorks/releases/tag/v6.1.0-rc12
I tried it and it worked for me. RC10 didn't work for me on Linux either.
@akenmorris RC12 worked for me as well
What's left?
I added more docs on the getting-started/shape to cover good and bad surfaces and what is shape modeling.
@akenmorris I added the tour video and the workshop videos (partitioned into three parts) to the docs. Nothing else for my side. Other than #1354, any thing else left for the release?
Thanks all for your great job in getting the 6.1 release out! Happy to hit the close button for this issue.
Please edit and add a ✅ indicating success and ❌ indicating failure or 🕒 for a test in progress with your username when you complete a task for a given platform.
When a test fails, please add a github issue, add it to the 6.1 release, and link it (*the issue when it's fixed and ready to test again). Also, go ahead and add new tasks that might not already be on here.
Please use the most recent release candidate for all testing (be careful which
shapeworks
is in your$PATH
). The most recent is RC14: https://github.com/SCIInstitute/ShapeWorks/releases/tag/v6.1.0-rc14Example:
Instructions for running and testing items
Testing modes:
General instructions
Notebooks
ShapeWorks/Examples/Python/notebooks/tutorials
on terminal/CMD.jupyter notebook
command.UseCases
ShapeWorks/Examples/Python/ on terminal/CMD
or on your favorite IDE that runs python.python RunUseCase.py --use_case <use_case>
command, wherepython RunUseCase.py --use_case femur
, orpython RunUseCase.py --use_case femur --groom_images
.Studio
ShapeWorks/Examples/Python/<item>
on terminal/CMD, where "ShapeWorksStudio <item>.xlsx
.Studio: MeshGrooming
ShapeWorksStudio
, or open ShapeWorksStudio somehow.Ok, now the real thing!
* Edit deep_ssm.py so that
num_samples = 3
on line 76 and"epochs": 3,
on line 132, otherwise it will take upwards of two days.Please keep comments on this issue to a minimum. Let's try to keep the status in the table and not in the comments.