Ethan Blackwood comments

Results 117 comments of


                                            Ethan Blackwood

Fix motion correction and CNMF for 3D movies

Caiman 1.11.2 is released

Kernel unexpectedly dying; line of code: viz_cnmf.show

@CharlotteWenxin did you get an error message like this in the Jupyter log? ``` thread '' panicked at src/conv.rs:1182:30: invalid size stack backtrace: 0: rust_begin_unwind ... ``` If so, I...

Kernel unexpectedly dying; line of code: viz_cnmf.show

@Agrajag1984 this issue is fixed in #309, we're waiting for a new caiman release to merge it but you should be able to use this branch for now: https://github.com/proektlab/mesmerize-core/tree/contour-nan-fix

Kernel unexpectedly dying; line of code: viz_cnmf.show

Ah yes I agree it's different from Charlotte's issue, but I believe this is caused by removing nans from empty contours (I got the same error with my data and...

Use Path.resolve() consistently to prevent UNC/drive letter ambiguity

@kushalkolar Here's an example of the problem, using some of the existing code from cnmf.py (sorry for the delay). The `resolve` method converts drive paths to UNC paths if the...

Use Path.resolve() consistently to prevent UNC/drive letter ambiguity

300 is merged, did you mean a different one?

Use Path.resolve() consistently to prevent UNC/drive letter ambiguity

Absolutely let me take a look later today! Fell through the cracks

Use Path.resolve() consistently to prevent UNC/drive letter ambiguity

@kushalkolar OK, I'm working on getting at least one of the test runs to succeed. Also, it seems like I still need an approving review from a reviewer with write...

Use Path.resolve() consistently to prevent UNC/drive letter ambiguity

I'm not reproducing this tolerance error on my Windows machine... `Max relative difference: 105.32658421` that seems really high, no? I see the rtol is already 100 though. @kushalkolar do you...

Use Path.resolve() consistently to prevent UNC/drive letter ambiguity

Anyway, the tests are all passing on my local setup with this issue, and the failure in the Windows runner here isn't related to the issue, so I'll go ahead...