Skip to content

Latest commit

 

History

History
146 lines (117 loc) · 5.51 KB

FAQs.md

File metadata and controls

146 lines (117 loc) · 5.51 KB

FAQs

This page answers some of the common problems encountered during installation of the 'meld_classifier' package and environment.

I have an error during the installation of the meld_classifier virtual environment for OS MAC1.

conda env create -f environment_MAC1.yml
  • if the error is : Solving environment: failed. ResolvePackageNotFound: tensorflow-deps

    You will need to force conda to look at osx-arm64:

    conda config --env --set subdir osx-arm64
    conda env create -f environment_MAC1.yml

I have an error during the installation of the meld_classifier virtual environment for Linux and other MAC OS.

conda env create -f environment.yml
  • if the error is : CondaValueError: prefix already exists: <path_to_anaconda>/anaconda3/envs/meld_classifier

    You need to remove the already existing environment and reinstall :

    rm -r <path_to_anaconda>/anaconda3/envs/meld_classifier
    conda env create -f environment.yml
  • if the error is : Your shell has not been properly configured to use ‘conda activate’

    First, ensure your conda path is in the '~/.bashrc' file and not in the '~/.bash_profile'.

    If in '~/.bash_profile', copy all the lines between >>> conda initialize >>> and <<< conda initialize <<< into the '~/.bashrc' file.

    nano ~/.bash_profile
    nano ~/.bashrc

    And add the following line just after:

    source <path_to_anaconda3>/etc/profile.d/conda.sh
    

    Save and close the terminal and open a new one.

  • if the error is : Your shell has not been properly configured to use ‘conda activate’

    You need to activate the bash shell:

    conda init bash

    Then close the terminal and open a new one.

  • if the error is : pip failed

    Open the environment.yml file and comment out all the line below the 1st 'pip' e.g:

    #- pip:
    #matplotlib_surface_plotting
    #- potpourri3d==0.0.4
    #- nilearn==0.8
    #- neuroCombat==0.2.12
    #- alibi==0.6
    #- patsy==0.5.2
    #- matplotlib==3.5.1
    #- torch==1.12
    #- torchvision==0.13
    #- SimpleITK==2.1.1
    #- git+https://github.com/Deep-MI/LaPy.git
    

    Re-install meld_classifier environment

    conda env create -f environment.yml

    And individually install the pip packages using :

    pip install <package_name>

    I have a package installation error when running the pytest.

    pytest
    • if the error is: Collected 0 items

      It might mean you have launched the pytest from the wrong folder. Ensure you are in the meld_classifier folder to run :

        cd <path_to_meld_classifier>
        conda activate meld_classifier
        pytest

      You should get Collected 45 items

I have an error when running the new patient pipeline scripts .

  • if the error happens after the predictions and is: FileNotFoundError: No such file or no access: .../meld_data/output/fs_outputs/<patient_id>/xhemi/surf_meld/rh.on_lh.thickness.mgh

    If this is the first time you are using the MELD pipeline and you have this error, you might have installed a version of Freesurfer that is too recent (V7.3 and above) and is incompatible with MELD. Check you Freesurfer version and if it is above V7.3 we would recommend deleting your current Freesurfer and/or installing Freesurfer V7.2 from https://surfer.nmr.mgh.harvard.edu/fswiki/rel7downloads and updating your paths to use this new version as shown in our guidelines and below:

    export FREESURFER_HOME=<freesurfer_installation_directory>/freesurfer
    source $FREESURFER_HOME/SetUpFreeSurfer.sh

    Then you will need to run the pipeline again from scratch, but before that you will need to ensure that you have deleted the folders and files below so that they can be recreated with the right version of Freesurfer:

    • all the folders and files in output/fs_outputs
    • the whole folder output/preprocessed_surf_data/MELD_<your_site_code>
      WARNING: it will delete all the outputs created by the MELD pipeline (except the predictions and reports), ensure that you don't need these data for other reasons
  • if the error is: It seems that scikit-image has not been built correctly. Your install of scikit-image appears to be broken.

    Reinstall scikit-image as follows:

      conda install -c conda-forge scikit-image=0.18

    And individually install llvm-openmp using :

    conda install -c conda-forge llvm-openmp
  • if the error concerned matplotlib

    Reinstall matplotlib as follows:

      pip uninstall matplotlib

    And individually install llvm-openmp using :

    pip install matplotlib
  • if the error looks like:

    dyld: lazy symbol binding failed: Symbol not found: ___emutls_get_address
    Referenced from: freesurfer/bin/../lib/gcc/lib/libgomp.1.dylib
    Expected in: /usr/lib/libSystem.B.dylib 
    dyld: Symbol not found: ___emutls_get_address
    Referenced from: freesurfer/bin/../lib/gcc/lib/libgomp.1.dylib
    Expected in: /usr/lib/libSystem.B.dylib
    

    It is an issue between Freesurfer V6 and the newer version of MacOs. This error can be solved by following the steps at the bottom of the Freesurfer web page

Contact the MELD team ([email protected]) for more information or help