search

DevelopmentalImagingMCRI / karawun

Convert tractography studies from mrtrix into dicom for use in navigation software
Apache License 2.0
14 stars 4 forks source link

PackagesNotFoundError: numpy==1.13.0 #12

Closed PatrickLYTang closed 3 years ago

PatrickLYTang commented 3 years ago

Hi,

When creating a conda environment I encounter a PackagesNotFoundError for numpy 1.13.0. Do you have any suggestions or experience with this error (in conda list numpy 1.17.0 is listed)?

Regards, Patrick

PackagesNotFoundError

richardbeare commented 3 years ago

The same command still works on a mac. I'll attempt the same test under windows in the office. At the moment I can only suggest looking for more channels - conda search numpy to see what shows up.

It probably means that it is time to bump versions of dependencies, possibly when the next SimpleITK is frozen.

richardbeare commented 3 years ago

Hi, I have a test version for you to try:

conda create --name KarawunEnv --file https://github.com/richardbeare/karawun/raw/NumpyDep/requirements.txt
conda activate KarawunEnv
pip install git+https://github.com/richardbeare/karawun.git@NumpyDep

If this does successfully install, I'd strongly recommending running the tests, as I've only done so under Linux.

Do the following with the KarawunEnv activated:

git clone https://github.com/richardbeare/karawun.git --branch NumpyDep
cd karawun
python -m pytest -s tests/

Hopefully they will all pass...

PatrickLYTang commented 3 years ago

Hi Richard,

Thanks a lot for the quick response. The installation of Karawun was successful! Unfortunately, parts of the tests raise an AssertionError. Do you have a suggestion what might be the cause?

Thanks in advance,

Patrick

image

richardbeare commented 3 years ago

I confirm the same failures under windows. This will take a little time to sort out. Hopefully it isn't anything dramatic. The files appear to get created, so you can have a play around importing to brainlab, but please don't do anything serious with the results for the moment.

richardbeare commented 3 years ago

Good news! The problem related to the test framework under windows (I was careless with paths, and the errors in the screenshot show different strings in the dictionary keys).

Add --upgrade --force-reinstall to your pip command, and retest. Alternatively you could do a "git pull" in the copy you cloned and install that. Let me know if it works and I'll merge into master.

PatrickLYTang commented 3 years ago

Hi Richard,

Thanks a lot! Indeed, there were no failures during the testing phase.

Regards,

Patrick