-
Notifications
You must be signed in to change notification settings - Fork 52
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
MAINT: Use PyBIDS 0.9.2+ #369
Conversation
I think this will require a 0.10.0 release to avoid version conflicts with things expecting pybids 0.7.x, templateflow 0.3.x, and niworkflows 0.9.x. |
It's not showing up yet, but I'm moving us to the PEP 440 Compatible Release syntax:
|
@oesteban The pure Python bits (I didn't build the container) test clean locally. Most of the changes are in the reports bit, so you should definitely have a look at that. |
Ah. Missed utils. Looks like there are some knock-on problems in templateflow, but I don't really know how to diagnose this:
|
Ah, looks like it's a bug in pybids. bids-standard/pybids#458 |
Okay, so I think I've fixed the bug in pybids... Now getting this one:
I think it's because MNIInfant doesn't get installed anywhere, which makes me wonder why it's not failing on Circle. |
Probably removing the templateflow folder will refresh the installation with the latest scaffold. |
Dockerfile
Outdated
python -c "from templateflow import api as tfapi; \ | ||
tfapi.get('MNI152Lin|MNI152NLin2009cAsym|OASIS30ANTs', suffix='T1w'); \ | ||
tfapi.get('MNI152Lin|MNI152NLin2009cAsym|OASIS30ANTs', desc='brain', suffix='mask'); \ | ||
tfapi.get('OASIS30ANTs', resolution=1, desc='4', suffix='dseg'); \ | ||
tfapi.get('OASIS30ANTs|NKI', resolution=1, label='brain', suffix='probseg'); \ | ||
tfapi.get('OASIS30ANTs|NKI', resolution=1, desc='BrainCerebellumRegistration', suffix='mask'); " | ||
tfapi.get('OASIS30ANTs|NKI', resolution=1, desc='BrainCerebellumRegistration', suffix='mask'); " && \ | ||
deactivate && rm -rf /tmp/venv |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
pybids was getting installed with templateflow, blocking its upgrade with requirements.txt
.
New issue:
Can't debug locally because I can't get templateflow configured properly. I can try the Docker image, but this is very annoying. |
Turns out it was templateflow/python-client#16. |
I agree. The most urgent thing, I believe, is making sure that updating TemplateFlow updates the HOME directory at installation (templateflow/python-client#11). Within the context of #374 we can address the problems of testing niworkflows. WDYT? |
Sounds reasonable. |
Now we're getting the MNIInfant thing on Circle, which is at least vaguely relieving because it's not fetched anywhere. It's unclear why it ever worked. |
@oesteban Now I can't replicate the failure locally. Is it just stochastic? I can try restarting, but it's a long wait to find out. |
It is very weird because a few lines above it says it is downloading There needs to be an obscure way that propagates some older templateflow folder. |
Hmm. It repeated, though the Docker image was identical. Do you see any ways forward? |
Maybe rerun the pytest build with ssh and check what the heck TemplateFlow
is using for home and whether stuff is really precached
…On Wed, Jul 10, 2019, 13:12 Chris Markiewicz ***@***.***> wrote:
Hmm. It repeated, though the Docker image was identical. Do you see any
ways forward?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#369>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAESDRUZUHIL47HVWUQ5JU3P6Y7DPANCNFSM4H7FXZVA>
.
|
Probably just easiest to ask: How would you identify the templateflow home and the status of the files? |
Maybe MNIInfant isn't always available via the S3 method? |
Sorry for the slow response. I believe I know what is happening. Please give me a second. |
Okay, yes. When TemplateFlow is not using DataLad, the first time the api is called the following happens:
What is happening? Updating the skeleton file is part of the packaging process. The problem is that, at this moment, the circleci build does not push back the new skeleton to the repo. Since you are installing from a commit, a very old skeleton is found in the repo (honestly, if it does not get updated it should go away). Fastest solution for you: cut an rc1 pre-release - it takes 5 minutes. |
Sure, let's do that then. Do you want to or should I? |
Thanks. |
Thank you for diving into these frozen waters |
Codecov Report
@@ Coverage Diff @@
## master #369 +/- ##
=========================================
Coverage ? 39.01%
=========================================
Files ? 43
Lines ? 4788
Branches ? 0
=========================================
Hits ? 1868
Misses ? 2920
Partials ? 0
Continue to review full report at Codecov.
|
Codecov Report
@@ Coverage Diff @@
## master #369 +/- ##
=========================================
Coverage ? 39.01%
=========================================
Files ? 43
Lines ? 4788
Branches ? 0
=========================================
Hits ? 1868
Misses ? 2920
Partials ? 0
Continue to review full report at Codecov.
|
This is a pre-condition for:
Depends on: