Installing a new version of mne seems to conflict with jupyter and spyder

I started by creating the conda virtual environment using anaconda.
conda create -n mne3 python=3.9
conda activate mne3

Then I installed jupyter and spyder.
conda install -c anaconda spyder
conda install -c conda-forge jupyterlab
And at this point, I started spyder and jupyter and everything was fine.

Then I installed mne, referring to the installation instructions for version 0.24, because the mamba installation method used in version 1.2 of mne does not work on my computer.

conda install -c conda-forge mne

Then after I installed mne, I started jupyer and spyder with the following error.
jupyter lab
10190 segmentation fault jupyter lab

/Users//opt/anaconda3/envs/mne3/bin/pythonw: line 3: 10217 Segmentation fault: 11 /Users///opt/anaconda3/envs/mne3/ “$@”

This is my system information
python -c “import mne; mne.sys_info()”
Platform: macOS-10.14.6-x86_64-i386-64bit
Python: 3.9.15 | packaged by conda-forge | (main, Nov 22 2022, 08:50:29) [Clang 14.0.6 ]
Executable: /Users//opt/anaconda3/envs/mne3/bin/python
CPU: i386: 8 cores
Memory: 16.0 GB
mne: 1.2.2
numpy: 1.21.6 {OpenBLAS 0.3.18 with 8 threads}
scipy: 1.9.3
matplotlib: 3.6.2 {backend=MacOSX}
sklearn: 1.0.2
numba: 0.55.1
nibabel: 4.0.2
nilearn: 0.9.2
dipy: 1.5.0
openmeeg: Not found
cupy: Not found
pandas: 1.5.2
pyvista: 0.37.0 {OpenGL 4.1 ATI-2.11.20 via AMD Radeon Pro 555 OpenGL Engine}
pyvistaqt: 0.9.0
ipyvtklink: 0.2.3
vtk: 9.1.0
qtpy: 2.2.0 {PyQt5=5.15.2}
ipympl: Not found
pyqtgraph: 0.13.1
pooch: v1.6.0
mne_bids: Not found
mne_nirs: Not found
mne_features: Not found
mne_qt_browser: 0.4.0
mne_connectivity: Not found
mne_icalabel: 0.4

if you only install mne does it work?

if you first install mne and then jupyterlab does it work?


1 Like

Hello @elvisjade, could you please elaborate on the issues you were experiencing with our mamba-based installation instructions? It’s very important to us to make it easy for users to install and use the latest versions of MNE-Python.

You could also consider using our standalone installers, which come with Spider and Jupyter Lab pre-installed!

Also as a general tip when installing packages with complex dependencies via conda or mamba: install eve thing in a single go, do not first install package A and then package B, but install all packages at once, with a single call to conda or mamba. This can avoid problems.

Best wishes,

1 Like

standalone installers might not be an option for me to consider.
Because I am more accustomed to using the conda environment, if there is any problem, I can delete the environment in time, which is more fault-tolerant for me.
If it is installed independently, it may involve problems such as the mac version or python in the mac system. If I do something wrong in some places, I may not be able to detect it myself.
Of course, independent installation is very convenient for many people, and I recommend it very much, but I prefer the conda environment.

1 if you only install mne does it work?

It is possible to install mne alone.
Whether you install mne first and then spyder jupyter or install spyder jupyter first, mne is always available.

2if you first install mne and then jupyterlab does it work?

Under this condition, jupyter is also unusable.

I guess it has something to do with mne using cpython? Because I was prompted to use cpython when installing mne.

This error is reported when I install using mamba.

mamba create --override-channels --channel=conda-forge --name=mne mne
ERROR transaction.cpp:84 File not valid: file size doesn't match expectation "/Users//opt/anaconda3/pkgs/click-8.1.3-py310h2ec42d9_1.tar.bz2"
click [======================] (00m:00s) Validating...
libc++abi.dylib: terminating with uncaught exception of type std::runtime_error: File not valid: file size doesn't match expectation (/Users/hasibagen/
click [======================] (00m:00s) Validating...
future [======================] (00m:00s) Validating...
[1] 10705 abort mamba create --override-channels --channel=conda-forge --name=mne mne

I googled this error. It seems that this error is because I used the mirror source of conda. Because I am in mainland China, the speed of directly accessing the source of conda is relatively slow. I used a mirror in China, but it seems to report an error.

This is my conda mirror source setup:

 conda info

     active environment : base
    active env location : /Users/hasibagen/opt/anaconda3
            shell level : 1
       user config file : /Users/hasibagen/.condarc
 populated config files : /Users/hasibagen/.condarc
          conda version : 4.12.0
    conda-build version : 3.21.5
         python version :
       virtual packages : __osx=10.14.6=0
       base environment : /Users/hasibagen/opt/anaconda3  (writable)
      conda av data dir : /Users/hasibagen/opt/anaconda3/etc/conda
  conda av metadata url : None
           channel URLs :
          package cache : /Users/hasibagen/opt/anaconda3/pkgs
       envs directories : /Users/hasibagen/opt/anaconda3/envs
               platform : osx-64
             user-agent : conda/4.12.0 requests/2.26.0 CPython/3.9.7 Darwin/18.7.0 OSX/10.14.6
                UID:GID : 501:20
             netrc file : None
           offline mode : False

The standalone installer simply deploys a conda environment and adds a few icons, that’s all. You end up with a complete and fully functional conda environment.

1 Like

Thank you for the explanation.
But this seems to require macOS 10.15 (Catalina) and newer, I’m on macOS 10.14.