lttbc icon indicating copy to clipboard operation
lttbc copied to clipboard

Issue with Numpy 1.22

Open ericnorway opened this issue 2 years ago • 7 comments

I ran into an issue with numpy 1.22.2 and lttbc 0.2.2. Downgrading to lttbc 0.2.1 or upgrading to numpy 1.24 solved the issue.

Exception has occurred: ImportError numpy.core.multiarray failed to import File "C:\temp\lttbc stuff\main.py", line 1, in import lttbc ImportError: numpy.core.multiarray failed to import

ericnorway avatar May 23 '23 11:05 ericnorway

thanks, I will have a look

dgoeries avatar May 23 '23 11:05 dgoeries

It looks like lttbc==0.2.2 is no longer in the pip registry. This is the response of the command

pip install lttbc==0.2.2

ERROR: Could not find a version that satisfies the requirement lttbc==0.2.2 (from versions: 0.1.9, 0.2.0, 0.2.1) ERROR: No matching distribution found for lttbc==0.2.2

vnadhan avatar May 23 '23 14:05 vnadhan

Yes, there is some trouble. I am sorry for that

dgoeries avatar May 23 '23 15:05 dgoeries

DO you have any updates here? When trying to install lttbc via poetry I run into a problem ModuleNotFoundError: No module named 'numpy' Poetry hints: Note: This error originates from the build backend, and is likely not a problem with poetry but with lttbc (0.2.1) not supporting PEP 517 builds. You can verify this by running 'pip wheel --use-pep517 "lttbc (==0.2.1)"'.

Is it possible that this error is also solved with the newer version.

I am using a Mac and lttbc don't have wheels for MacOS.

eyjay-ok avatar Jun 08 '23 06:06 eyjay-ok

There will be a new version fixing this. I had a version up that unfortunately was always taken the latest numpy available when specified in the build requirements, providing incompatibilities.

Will try again and then upload a new version, sorry for inconvenience

dgoeries avatar Jun 08 '23 08:06 dgoeries

Running through a lot of documents I could not fully find a satisfying solution. I went for pinning numpy 1.22 in the build process. Specifying a build requirement with >= always did lead to the latest being used. This resulted in incompatible versions. Googling and looking around, it looks like this package is not the only one suffering from this. Numpy has a large backward compatibility ...

Due to Python 3.11 support I went for 1.22 as requirement.

dgoeries avatar Jun 10 '23 12:06 dgoeries

The new version works for me. Thanks :)

eyjay-ok avatar Jun 12 '23 11:06 eyjay-ok