How do you install pyiron_nodes?

[XzzX]

How do you install pyiron_nodes? I tried a few ways but always got stuck somewhere.

Not found in conda-forge, compilation issues when installing with pip, deprecated not in pyiron_base when installing everything myself, etc, etc

[liamhuber]

It's just git-clone. It's not currently set up for auto-publishing. This is intentional while the repo currently serves as a public but pseudo-internal grab bag of various only-loosely-related applications. @pyiron/distribution might have more up-to-date details on the plan moving forward.

[XzzX]

With git clone alone I am missing dependencies. And pip install . exits with

...
      Failed to build pyscal3
      ERROR: ERROR: Failed to build installable wheels for some pyproject.toml based projects (pyscal3)
      [end of output]
  
  note: This error originates from a subprocess, and is likely not a problem with pip.
error: subprocess-exited-with-error

× pip subprocess to install build dependencies did not run successfully.
│ exit code: 1
╰─> See above for output.

note: This error originates from a subprocess, and is likely not a problem with pip.

I need it to reproduce Joergs tests with my storage/database interface.

[liamhuber]

The dependencies are listed in .ci_support/environment.yml, all with specific versions. Once you've cloned the repo, it is only a short step to create a conda environment from this file. I don't know why pip complains about pyscal, but the tests build the conda env so I expect that route to work.

Of course, I can only guarantee that this is the right environment insofar as the various contributors have their contributions included in the test suite and have duly configured that file so their work runs.

In principle, getting it pip-installable should be as easy as changing the flag I previously linked from false to true, and then tagging a github release. Making the package available on conda-forge is then slightly more work, but also very achievable. However, as far as I understand, not doing these steps is completely intentional at present. If you want this repo to be available on pip and conda, you need buy-in from @pyiron/distribution. Personally, I would advise against publication while the repo content is still so fragile (very little test coverage) and poorly documented.

[XzzX]

The dependencies are listed in .ci_support/environment.yml, all with specific versions. Once you've cloned the repo, it is only a short step to create a conda environment from this file. I don't know why pip complains about pyscal, but the tests build the conda env so I expect that route to work.

Of course, I can only guarantee that this is the right environment insofar as the various contributors have their contributions included in the test suite and have duly configured that file so their work runs.

In principle, getting it pip-installable should be as easy as changing the flag I previously linked from false to true, and then tagging a github release. Making the package available on conda-forge is then slightly more work, but also very achievable. However, as far as I understand, not doing these steps is completely intentional at present. If you want this repo to be available on pip and conda, you need buy-in from @pyiron/distribution. Personally, I would advise against publication while the repo content is still so fragile (very little test coverage) and poorly documented.

Thx, seems to work. But pyiron_nodes does not work with the default pyiron_base version.

ImportError: cannot import name 'deprecate' from 'pyiron_base'

[liamhuber]

I'm on mobile so I can't easily dig into where the error is coming from -- but wherever it is, the reference needs to be changed to the pyiron_snippets module, which got the deprecation tools from base some time ago.