Closed jmdelahanty closed 1 year ago
Previously suite2p depended on scanimage-tiff-reader, but it was no longer supported on pypi. Thanks to Lawrence Niu (MBF Bioscience engineer) this is now fixed! Please pip install scanimage-tiff-reader
and see if that removes the warnings. I think it may not install on Mac still, but it's working on Windows and Linux.
This worked! Errors are no more.
The newest version of suite2p (suite2p v0.13.0) is spitting out many
tifffile
messages while attempting to run on a dataset. None of my labmates have seen this behavior yet and I haven't seen any issues related to it here. It doesn't seem to be a bug since it appears that things are still running normally, but I get many error messages logged like this one:I think it's doing this for every image in the recording it attempts to load. Examining an individual tiff like this using tifffile installed in my suite2p environment:
Gives me the same error, like this:
Looking through a couple different recordings shows that Bruker's .ome.tif output simply puts a value of
[0]
in the time created tag.I found a similar issue elsewhere describing this as a fatal error in the past, but Christophe Gohlke resolved it. Funny enough it references PrairieView doing something odd as the reason it was raised in the first place.
A solution for avoiding logging these kinds of things could be just to ignore these types of warnings, but I don't know what best practices are or if its even worth it. Not sure why I or my labmates haven't encountered this using previous versions of suite2p either.
EDIT ~I think this could possibly impact performance, it at least "feels" like its taking longer than normal on a machine I know is quite speedy for getting things into binary. Once I have some numbers comparing a dataset with these errors vs an instance where I don't I'll post them here...~
It does not influence performance.