Python modules: separate import-time definitions from program execution

Operation contract

The program defines a receipt-label function and guards the printing entrypoint with __name__ == "__main__". A caller importing that source can reuse the function without printing a receipt immediately. Importing a module still executes its definitions and other unguarded top-level work; the guard does not make every import side-effect free.

Failure and ownership boundary

A local file named json.py or logging.py can shadow an intended module on the search path. Put reusable code in an owned package and keep environment-dependent network or database setup out of import-time code. Python interpreter and virtual environments: run the intended executable and Python receipt CLI project: parse options and report failed input make that execution boundary explicit.

Working program

def receipt_label(identifier):
    if identifier <= 0:
        raise ValueError("positive identifier required")
    return f"receipt:{identifier}"

if __name__ == "__main__":
    print(receipt_label(41))

Output

receipt:41

Costs and limits

The small function allocates one label string. Import cost includes module initialization and transitive imports, which this single-file fixture does not benchmark.

Common Mistakes

• A main guard does not erase unguarded import work.
• Do not shadow library module names with local files.

Connected lessons

Python interpreter and virtual environments: run the intended executable, Python receipt CLI project: parse options and report failed input, Java modules: readability, exports and reflection boundaries.

Apply this boundary

Python package layout: owned imports and module entrypoints, Python process pools: importable workers and serialization costs.

Check the next state boundary

Python import failures: module removal does not roll back every side effect, Python plugin project: load owned code and publish only an accepted interface.

Check this related boundary

Python module cache: repeated imports reuse an owned module object.