setup

Check the current Python environment for libraries required by the ds plugin. Report what is installed, what is missing, and provide install commands.

Steps

1. Check Python version

Run python3 --version and report the result. Require Python 3.9+.

2. Check required libraries

For each required library, run a Python import and version check:

python3 -c "
import importlib
required = {
    'pandas': 'pandas',
    'scikit-learn': 'sklearn',
    'scipy': 'scipy',
    'statsmodels': 'statsmodels',
    'numpy': 'numpy',
}
for pkg_name, import_name in required.items():
    try:
        mod = importlib.import_module(import_name)
        version = getattr(mod, '__version__', 'installed')
        print(f'  {pkg_name}: {version}')
    except ImportError:
        print(f'  {pkg_name}: MISSING')
"

3. Check optional libraries

Run the same check for optional libraries used in generated experiment code:

python3 -c "
import importlib
optional = {
    'matplotlib': 'matplotlib',
    'seaborn': 'seaborn',
    'aeon': 'aeon',
    'xgboost': 'xgboost',
    'lightgbm': 'lightgbm',
    'shap': 'shap',
    'great_expectations': 'great_expectations',
    'polars': 'polars',
    'optuna': 'optuna',
}
for pkg_name, import_name in optional.items():
    try:
        mod = importlib.import_module(import_name)
        version = getattr(mod, '__version__', 'installed')
        print(f'  {pkg_name}: {version}')
    except ImportError:
        print(f'  {pkg_name}: not installed')
"

4. Report install commands

If any required libraries are missing, output the install command:

uv pip install pandas scikit-learn scipy statsmodels numpy

List only the missing packages in the command. If all required libraries are present, report that the environment is ready.

For missing optional libraries, suggest but do not insist:

uv pip install matplotlib seaborn aeon xgboost lightgbm shap great_expectations polars optuna

Rules

• Do NOT auto-install anything. Data scientists manage their own environments.
• Do NOT suggest installing into the system Python. Assume the user has an active virtual environment or conda environment.
• If python3 is not found, suggest the user activate their environment first.
• Report results as a simple table, not verbose prose.