Skip to content
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

Set gram_schmidt tolerances in POD to 0 to never truncate pod modes #980

Merged
merged 2 commits into from
Jul 7, 2020

Conversation

sdrave
Copy link
Member

@sdrave sdrave commented Jul 2, 2020

Fixes #961.

@sdrave sdrave added this to the 2020.1 milestone Jul 2, 2020
@sdrave sdrave added the pr:fix Fixes a bug label Jul 2, 2020
@sdrave sdrave requested a review from pmli July 2, 2020 16:30
@sdrave
Copy link
Member Author

sdrave commented Jul 3, 2020

@renefritze, here we have another failed build that has no hash for reproducing and very long error message.

@renefritze
Copy link
Member

That actually looks like the log was cut. I'm making the hypothesis dir downloadable to better debug cases like this. The new push might just clear the error too though.

@renefritze
Copy link
Member

@sdrave
Copy link
Member Author

sdrave commented Jul 6, 2020

@renefritze, I can't reproduce the failures locally. Can you?

Also, why are there so many 'Trying example:' blocks? I still don't really understand the output. There are also two different 'reproduce_failure' decorators mentioned. Why?

@renefritze
Copy link
Member

I can't reproduce the failures locally.

Even using the right docker image?

Also, why are there so many 'Trying example:' blocks?

An "example" is one input state for the test function. max_examples is 100 for pull requests.

There are also two different 'reproduce_failure' decorators mentioned. Why?

Cause it found multiple failures.

@sdrave
Copy link
Member Author

sdrave commented Jul 6, 2020

I can't reproduce the failures locally.

Even using the right docker image?

Yes.

Also, why are there so many 'Trying example:' blocks?

An "example" is one input state for the test function. max_examples is 100 for pull requests.

But why do I see the output for all these examples? Can this be turned of?

There are also two different 'reproduce_failure' decorators mentioned. Why?

Cause it found multiple failures.

I would have assumed that it stops when it finds a failing example. Is the second one the 'reduced' version?

@renefritze
Copy link
Member

But why do I see the output for all these examples?

I think you'll have to ask the library authors for this.

Can this be turned of?

Haven't seen a setting that controls output, but I haven't looked for one.

I would have assumed that it stops when it finds a failing example. Is the second one the 'reduced' version?

E hypothesis.errors.MultipleFailures: Hypothesis found 2 distinct failures.

You can make hypothesis stop after it found one though, but I don't think that's a good choice for CI.

@sdrave
Copy link
Member Author

sdrave commented Jul 6, 2020

Can this be turned of?

Haven't seen a setting that controls output, but I haven't looked for one.

In conftext.py you have set the verbosity of hyporthesis to verbose. If I understand the code correctly, setting it to normal should print only failing examples.

@sdrave sdrave mentioned this pull request Jul 6, 2020
@sdrave
Copy link
Member Author

sdrave commented Jul 7, 2020

So after rebasing on master, the hypothesis failure is gone, but I suspect we will see it again at some point .. - @renefritze should this be merged with 7e869c3 still in it?

@codecov
Copy link

codecov bot commented Jul 7, 2020

Codecov Report

Merging #980 into master will increase coverage by 0.03%.
The diff coverage is n/a.

Impacted Files Coverage Δ
src/pymor/vectorarrays/numpy.py 84.31% <0.00%> (-0.25%) ⬇️
src/pymor/bindings/ngsolve.py 85.29% <0.00%> (+0.73%) ⬆️
src/pymor/vectorarrays/list.py 84.33% <0.00%> (+0.94%) ⬆️

@renefritze
Copy link
Member

So after rebasing on master, the hypothesis failure is gone, but I suspect we will see it again at some point .. - @renefritze should this be merged with 7e869c3 still in it?

Yes. That doesn't hurt and can be useful in a future occurrence.

@sdrave sdrave merged commit 87201fb into master Jul 7, 2020
@sdrave sdrave deleted the fix_961_no_truncation branch July 7, 2020 16:09
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
pr:fix Fixes a bug
Projects
None yet
Development

Successfully merging this pull request may close these issues.

POD singular values not truncated when gram_schmidt removes vectors
3 participants