Execution failed: Could not connect to the Python process.

Hello all!

I am trying to run a script that calculates the centered log ratio.

What makes my situation potentially different, from past questions on the same topic is that the Python Scrip Node used to work flawlessly with the same script. But now as the screenshot shows, the Python Node fails to work. Im not sure what is different now than in the past. I do have installed the basic Python extension.

Here is the flow variables which I have not changed. . . .

image

By training I am a geoscientist. . . . Im taking baby steps into data science by teaching myself. I hope someone can help me.

Many thanks!

~Cole K.


Here is some additional information. Im using the standard environment which worked in the past.

Hi Cole. Are you using macOS 15.4? There is a known issue with that in general, more information can be found here - Numpy import error when using conda-forge build of numpy 2.2.3 // no issue with pip · Issue #347 · conda-forge/numpy-feedstock · GitHub

Unfortunately that causes issues with Python nodes in Knime as well. We have a script that was created by our engineering team to have a workaround - but that is currently being tested.

Would you be open to helping us test this workaround on your side? While a permanent fix will require waiting for the related GitHub issue to be resolved and for us to release a new bundled environment, this temporary solution might help unblock you in the meantime.

Let us know if you’re able to give it a try!

2 Likes

Hi Bence! I did update to MacOS 15.4 so that may actually be the problem. I would be pleased to act as a beta tester for this work-around that you state. I would simply need some basic-level instructions (that a non-specialist like me would easily understand) into how to implement it on my end.

Thank you again for getting back to me! Also please keep me updated into how the permanent solution is coming.

~Cole K.

Hi Bence! I did update to MacOS 15.4 so that may actually be the problem. I would be pleased to act as a beta tester for this work-around that you state. I would simply need some basic-level instructions (that a non-specialist like me would easily understand) into how to implement it on my end.

Thank you again for getting back to me! Also please keep me updated into how the permanent solution is coming.

~Cole K.

Hi Cole,

He is a link to download - Box

(only available for a few days)

The script can be used to fix all Python environments bundled with a KNIME installation:
./fix_knime_python_envs.sh /Applications/path/to/KNIME.app

Or to fix an external conda environment:
./fix_knime_python_envs.sh /path/to/conda/env

What you need to do is

  1. Open terminal and navigate to the script’s location
  2. Make the script executable
    chmod +x fix_knime_python_envs.sh
  3. Find the correct path to KNIME.app
    ( You may need to escape spaces using backslashes (\ ) or quote the entire path:) Example:
    ./fix_knime_python_envs.sh /Applications/KNIME\ 5.4.3.app
  4. If you get a permission denied error run with sudo but in general I would recommend asking ChatGPT what it recommends doing about the error you got.

If everything is correct it will start checking libraries in the Knime installation.

Let me know if that worked.

1 Like

Hello Bence! I am pleased to announce here today that your solution, the download that you provided in tandem with help from Sonnet 3.7 worked seamlessly together to single-handedly and completely reverse the stress and agony that yesterday afternoon provided me. The python node in KNIME now works beautifully and flawlessly. I am forever grateful for the assistance you provided me. So Grateful! Would recommend framing.

Have a good rest of the day and the upcoming weekend!

~Cole K.

Great to hear it worked. Here you can find the official forum post about the issue by our developers:

2 Likes

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.