commands/sysadmin/linux-desktop/python-environments/manage-conda-environments.md

Manage Conda Environments

You are helping the user list conda environments and work with them to add packages.

Your tasks:

1. Check if conda is installed:

   which conda
   conda --version
   conda info
   

   If not installed, offer to help install Miniconda or Anaconda.

2. List all conda environments:

   conda env list
   # or
   conda info --envs
   

   This shows:

   • All environment names
   • Their locations
   • Current active environment (marked with *)

3. Show current environment:

   echo $CONDA_DEFAULT_ENV
   conda info --envs | grep "*"
   

4. Display detailed environment information: For each environment, show:

   # List packages in specific environment
   conda list -n <env-name>
   
   # Show environment details
   conda env export -n <env-name>
   
   # Show size
   du -sh ~/miniconda3/envs/<env-name>
   # or
   du -sh ~/anaconda3/envs/<env-name>
   

5. Ask user which environment to work with: Present the list and ask which environment they want to modify or examine.

6. Activate environment:

   conda activate <env-name>
   

   Verify activation:

   conda info --envs
   python --version
   which python
   

7. Show packages in environment:

   conda list
   # or for specific environment
   conda list -n <env-name>
   
   # Show only explicitly installed packages
   conda env export --from-history -n <env-name>
   

8. Search for packages: Ask what packages user wants to install:

   conda search <package-name>
   conda search <package-name> --info
   

9. Install packages:

   Single package:

   conda install <package-name>
   # or specify environment
   conda install -n <env-name> <package-name>
   

   Multiple packages:

   conda install <package1> <package2> <package3>
   

   Specific version:

   conda install <package-name>=<version>
   # Example:
   conda install python=3.11
   conda install numpy=1.24.0
   

   From specific channel:

   conda install -c conda-forge <package-name>
   

10. Suggest common packages by category:

    Data Science:

    conda install numpy pandas matplotlib seaborn scikit-learn
    conda install jupyter jupyterlab notebook
    conda install scipy statsmodels
    

    Machine Learning:

    conda install tensorflow pytorch torchvision
    conda install keras scikit-learn xgboost
    conda install -c conda-forge lightgbm
    

    Development:

    conda install ipython black flake8 pytest
    conda install requests beautifulsoup4 selenium
    conda install flask django fastapi
    

    Visualization:

    conda install matplotlib seaborn plotly
    conda install bokeh altair
    

    Database:

    conda install sqlalchemy psycopg2 pymongo
    conda install sqlite
    

11. Update packages:

    Update specific package:

    conda update <package-name>
    

    Update all packages in environment:

    conda update --all
    

    Update conda itself:

    conda update conda
    

12. Remove packages:

    conda remove <package-name>
    # or from specific environment
    conda remove -n <env-name> <package-name>
    

13. Create new environment: Offer to create a new environment:

    # Basic environment
    conda create -n <env-name> python=3.11
    
    # With packages
    conda create -n myenv python=3.11 numpy pandas jupyter
    
    # From file
    conda env create -f environment.yml
    

14. Export environment: Help user export environment for sharing:

    Full export (with all dependencies):

    conda env export -n <env-name> > environment.yml
    

    Only explicitly installed packages:

    conda env export --from-history -n <env-name> > environment.yml
    

    As requirements.txt:

    conda list -n <env-name> --export > requirements.txt
    

15. Clone environment:

    conda create --name <new-env> --clone <existing-env>
    

16. Clean up conda:

    # Remove unused packages and caches
    conda clean --all
    
    # Remove packages cache
    conda clean --packages
    
    # Remove tarballs
    conda clean --tarballs
    
    # Check what would be removed
    conda clean --all --dry-run
    

17. Check environment conflicts:

    # Check for broken dependencies
    conda info <package-name>
    
    # Verify environment
    conda env export -n <env-name> | conda env create -n test-env -f -
    

18. Show environment size:

    # Size of all environments
    du -sh ~/miniconda3/envs/*
    # or
    du -sh ~/anaconda3/envs/*
    
    # Total conda size
    du -sh ~/miniconda3
    

19. Configure conda:

    Add channels:

    conda config --add channels conda-forge
    conda config --add channels bioconda
    

    Set channel priority:

    conda config --set channel_priority strict
    

    Show configuration:

    conda config --show
    conda config --show channels
    

    Remove channel:

    conda config --remove channels <channel-name>
    

20. Use mamba (faster alternative): If user has performance issues:

    # Install mamba
    conda install mamba -n base -c conda-forge
    
    # Use mamba instead of conda
    mamba install <package-name>
    mamba search <package-name>
    mamba env create -f environment.yml
    

21. Troubleshooting common issues:

    Environment not activating:

    conda init bash
    source ~/.bashrc
    

    Package conflicts:

    # Create new environment instead
    conda create -n new-env python=3.11 <packages>
    

    Slow package resolution:

    # Use mamba
    conda install mamba -c conda-forge
    # or
    conda config --set solver libmamba
    

    Conda command not found:

    export PATH="$HOME/miniconda3/bin:$PATH"
    conda init bash
    

22. Best practices to share:

    • Create separate environment for each project
    • Use environment.yml for reproducibility
    • Pin important package versions
    • Use conda-forge channel for latest packages
    • Regularly clean up with conda clean --all
    • Don't install packages in base environment
    • Use mamba for faster package resolution
    • Export environments before major changes
    • Keep Python version explicit in environment
    • Use --from-history for cross-platform compatibility

23. Show workflow example:

    # Create environment
    conda create -n data-project python=3.11
    
    # Activate it
    conda activate data-project
    
    # Install packages
    conda install numpy pandas jupyter matplotlib scikit-learn
    
    # Verify
    conda list
    
    # Export for sharing
    conda env export --from-history > environment.yml
    
    # Deactivate when done
    conda deactivate
    

24. Integration with Jupyter:

    # Install ipykernel in environment
    conda activate <env-name>
    conda install ipykernel
    
    # Register environment as Jupyter kernel
    python -m ipykernel install --user --name=<env-name>
    
    # Now available in Jupyter
    jupyter lab
    

25. Report current status: Summarize:

    • Number of environments
    • Active environment
    • Total disk usage
    • conda version
    • Suggested actions based on user needs

Important notes:

• Always activate environment before installing packages
• Base environment should be kept minimal
• Use -n <env-name> to work with environments without activating
• conda-forge channel has more packages than default
• mamba is drop-in replacement, much faster
• Environment.yml files ensure reproducibility
• Pin versions in production environments
• Clean up regularly to save disk space
• Don't mix pip and conda unless necessary (prefer conda)
• Use --from-history when exporting for other OS
• Jupyter needs ipykernel in each environment
• conda init modifies .bashrc - check if needed