Hi @JoOkuma, your software looks awesome! I am a maintainer of PlantSeg and @tischi is trying to use it for one of our project. I had a look at #84 and made a few fixes. It looks like contributions are welcomed, so I'll just make PR here:
In the environment config in PlantSeg example, I changed the name and channel of PlantSeg, since we moved our software to conda-forge channel already. I also made all dependencies in these files pure conda-based, as they provides the same packages. Specifying a version for pytorch-cuda is recommended by PyTorch and cuCIM installation.
In the main README and the Q&A in doc, I emphasis the need of installing pyqt via -c conda-forge prior the installation of ultrack, since Napari is a default dependency and it requires but doesn't come with Qt. Note that pyqt from conda-forge channel is the most robust installation from my experience and should guarantee the launch of Napari.
I tested the installation on EMBL cluster and my system to run PlantSeg Napari pipeline and Ultrack tracking on @tischi's dataset without a problem, so I believe it's ready to be merged. Please have a look and see if things need to be modified.
Hi @JoOkuma, your software looks awesome! I am a maintainer of PlantSeg and @tischi is trying to use it for one of our project. I had a look at #84 and made a few fixes. It looks like contributions are welcomed, so I'll just make PR here:
conda-forge
channel already. I also made all dependencies in these files pure conda-based, as they provides the same packages. Specifying a version forpytorch-cuda
is recommended by PyTorch and cuCIM installation.pyqt
via-c conda-forge
prior the installation ofultrack
, since Napari is a default dependency and it requires but doesn't come with Qt. Note thatpyqt
fromconda-forge
channel is the most robust installation from my experience and should guarantee the launch of Napari.I tested the installation on EMBL cluster and my system to run PlantSeg Napari pipeline and Ultrack tracking on @tischi's dataset without a problem, so I believe it's ready to be merged. Please have a look and see if things need to be modified.