-
Notifications
You must be signed in to change notification settings - Fork 176
Fix point evaluation API change #4675
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
Conversation
|
This is a bug fix (and a critical one at that) so should go into |
|
We should also fix calling a We need something like |
|
Ideally we would want |
2df11b3 to
b95a4a0
Compare
f94ae63 to
af5cfcb
Compare
|
@leo-collins the recommendations from the meeting yesterday were:
I confess I got a bit confused by what we ended up deciding to do but if you agree with that please implement it here and poke me for a review when the tests are passing. I can then make a new release. |
fix more test tensor tests fix
lint
5551e22 to
53eaec4
Compare
|
@connorjward This should be okay to merge (assuming the tests pass) |
|
@leo-collins I am happy but please can you put more information in the PR description explaining all this? |
|
@connorjward Updated description. |
Description
I accidentally changed the API of
Function.evaluatein #4543. This changes the shape of the output to how it was before. Now, if you pass a single point toFunction.evaluate, then we remove the first axis of the result. For example, ifthen
u((0.5, 0.5))returns a numpy array with shape(), and not an array with shape(1,). Ifulives in aVectorFunctionSpace, thenu((0.5, 0.5))returns an array with shape(2,), and not an array with shape(1,2).There were no tests looking for this before so I've added some.