Clean Python Code

Personalized Python and PySpark formatting for VS Code, based on a hierarchical indentation style for analytics code.

Why Clean Python Code?

Clean Python Code is designed for developers and data professionals who want more control over how complex Python/PySpark expressions are laid out.

It focuses on visual hierarchy, not on enforcing Black or conventional PEP 8 formatting.

Main command

Select a Python code block and press:

Ctrl + I

The shortcut is intentionally registered as ctrl+i on Windows, Linux, and macOS.

What it formats

• Chained PySpark/DataFrame operations such as .filter(), .withColumn(), .select(), .groupBy(), .agg(), .join() and other method chains.
• Long one-line expressions by introducing aligned parentheses.
• Long argument lists, lists, tuples and dictionaries.
• Nested calls such as F.when(...), isin(...), aggregations and Window expressions.
• Complex boolean expressions using &, |, and and or.
• Python control-flow structures such as for, if, elif, else, try, except, finally, with, match, functions and classes, while preserving Python's structural indentation.
• Multiple statements, blank lines and comments.
• Python cells inside .ipynb notebooks opened in VS Code/Jupyter.

Reference style

The formatter follows the supplied reference examples, including leading commas in expanded collections and function arguments.

Before

df = spark.table('database.dbo.my_table').filter(F.col('system_id') == '123#').filter(F.col('dt') > '2026-06-01').select(F.col('system_id'), F.col('client_id'))

After

df = (
        spark.table('database.dbo.my_table')
             .filter(F.col('system_id') == '123#')
             .filter(F.col('dt') > '2026-06-01')
             .select(
                      F.col('system_id')
                      ,F.col('client_id')
                    )
     )

Complex nested expression

.withColumn(
    "risk_level",
    F.when(
        (F.col('score') >= 0.8)
        & (F.col('status') == 'ACTIVE'),
        "HIGH"
    ).otherwise("LOW")
)

Collection formatting

accounts = [
              "abc##"
              ,"dfo..."
              ,"12009c"
           ]

Settings

All options are available under Settings → Clean Python Code:

Supported files

• .py
• Python cells in .ipynb notebooks supported by VS Code/Jupyter

Design principles

Clean Python Code intentionally does not attempt to replace Black, Ruff, autopep8 or import sorters. The formatter is opinionated about visual structure and is intended to complement normal Python tooling.

The formatter is designed to be idempotent: applying Ctrl + I repeatedly should not keep changing the result.

Installation from VSIX

For local testing:

code --install-extension clean-python-code-0.3.0.vsix

Or use Extensions → Views and More Actions (...) → Install from VSIX....

Marketplace

Clean Python Code is published under the publisher associated with the extension manifest. The package uses the Marketplace extension id:

cleanpythoncode.clean-python-code

Replace cleanpythoncode in package.json with your actual Marketplace Publisher ID before the first publication if you use a different publisher.

Support

For bug reports and feature requests, open an issue in the project's public repository.