DPU installation issue - Exit code 1

Hi all,

We have purchased a new Windows computer for the eVOLVER and are reinstalling the software. I ran into this issue:

The virtual environment is running python 3.11 and both pip and setuptools have been checked for upgrades.

Any suggestions?

Hi @karolbuda, I think you just need to cd into the dpu directory.

Oh nevermind I see you have renamed your directory. Let me dig into it and get back to you!

Try running:

pip install -r requirements.txt

I think some new updates for setuptools might have broken something with setup.py.

Brandon

I just confirmed that trying to install using python10 on my mac will not install bokeh, but it succeeds using python8. We should try to upgrade bokeh to the latest version asap. Not sure if this is the same issue @karolbuda is having but it’s the point that breaks for me on my machine. Hopefully Brandons suggestion above works for you!

Zack

Thanks both for the prompt replies as always! I’ll test this out (as well as alternative python versions) on Monday and let you know how it goes.

Hei there @heinsz and @bgwong,
has this error been fixed? Having the exact same problem right now - Windows machine, python 3.11 installed, running the pip install -r requirements.txt works like a charm, but the python -m pip install . results in the same error.
Any (new) suggestions?

Edit:
Not exactly the same, the error is:

 Installing build dependencies ... done
  Getting requirements to build wheel ... error
  error: subprocess-exited-with-error

  × Getting requirements to build wheel did not run successfully.
  │ exit code: 1
  ╰─> [15 lines of output]
      Warning: 'classifiers' should be a list, got type 'tuple'
      error: Multiple top-level packages discovered in a flat-layout: ['graphing', 'experiment', 'calibration'].

      To avoid accidental inclusion of unwanted files or directories,
      setuptools will not proceed with this build.

      If you are trying to create a single distribution with multiple packages
      on purpose, you should not rely on automatic discovery.
      Instead, consider the following options:

      1. set up custom discovery (`find` directive with `include` or `exclude`)
      2. use a `src-layout`
      3. explicitly set `py_modules` or `packages` with a list of names

      To find more information, look for "package discovery" on setuptools docs.
      [end of output]

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

× Getting requirements to build wheel 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.

Hi there,

We haven’t updated everything to use python 3.11 - this could be the cause of the issues. The last fresh install I did was using python 3.8 on a windows machine and it worked. It’s on the radar to update and move to a better dependency management tool like poetry with toml files.

Zack

I just had this same issue on my Mac (M2, Ventura 13.2.1) with python 3.9.6 installed as default. I used pyenv to revert to 3.8.16 and the installation appears to have worked.

2 Likes

Interesting. I did try to remove (left-click, uninstall) the different python packages from the windows machine and installing another version via the Microsoft App Store; python --version confirms the correct version. None of the versions did work.
It always gives the same error
Warning: 'classifiers' should be a list, got type 'tuple' error: Multiple top-level packages discovered in a flat-layout: ['graphing', 'experiment', 'calibration'].
I was trying to start to install some packages manually in the virtual environment, based on the error messages I get from the GUI, and that worked to some extend, but it seems to be like a Sisyphos task :smiley:

Tried with a Mac now… Same error. I did also try multiple python versions on windows and on mac, always that error.

Any other suggestions @heinsz @bgwong ? :smiley:

Running the requirements.txt works however…

Sorry for the delayed response - you were able to install everything?

Hei,
No worries :smiley:
Well, running the VM, specifically running the installation script still fails, same error. This happened on a new/clean installation on Mac and on Windows as well, I tried multiple python versions. Maybe you could just change the (…) to a […] for the classifiers, if non specific tuple features are needed?

I can start the evolver and everything, but now I ran into some connection problems, specifically that some IO ports are already in use or have been closed already. Have the suspicion that might be due to the network though - I hope that at least :wink:

Hey Christian,

I got around this problem by not upgrading my pip during installation. So delete your dpu-env folder and do the same dpu-env process without upgrading pip. My mac had pip 20.2.3 and it gave me warnings, but still did the installation

Best,
Nate

1 Like

Brilliant, that fixed the problem! Thanks

A user has a new Mac that comes with python 3.11.4, which had the same errors even with the hack of having pip 20.2.3 installed.

We downgraded python to 3.8.17 using pyenv and then also pip to 20.2.3 to get things to install.

So seems like we need a more permanent solution for this because the DPU installation is failing.

@heinsz what should we do? Last time we talked we mentioned switching to conda environments and having users install miniconda