Check out pyimagej 0.4.0

With the 0.4.0 release of the ImageJ Python module—renamed from imagej to pyimagej now—it is feasible to start up an ImageJ without needing a local copy of Fiji:

import imagej
ij = imagej.init()

By default, it grabs the latest release of the net.imagej:imagej artifact from the maven.imagej.net.

You can also spin up a different version explicitly for reproducibility, or spin up an ImageJ1-enabled ImageJ, or a full Fiji installation. See the README for details.

One of these capabilities may help streamline relevant ImageJ usage in example notebooks here.

I did some testing but there are surely still some rough edges. If you get a chance to play with it and find any issues, please do not hesitate to let me know.

The release is available now on PyPI, and should be on conda-forge as well shortly.

[ctrueden]

Closing this issue here in favor of imagej/pyimagej#21 and scijava/scyjava#5. Once those are resolved, and I’m confident pyimagej is usable in a dynamic notebook context via ! syntax, we can try again to update the notebooks here.

[thewtex]

I agree that would be a nice improvement. Could you open an issue on the pyimagej repo?

Yes! Added with a minimal example here: https://github.com/imagej/pyimagej/issues/21