pycommremover

It analyses the code with the standard-library tokenize and ast modules, so a # inside a string is never mistaken for a comment and real string values (assignments, call arguments, f-strings) are never touched. Line numbers of the remaining code are preserved.

Why three functions?

At the language level these are different things, so they are removed separately:

remove_comments is behaviour-preserving (it only deletes no-ops and keeps docstrings). remove_docstrings changes __doc__, so it is opt-in. remove_comments_and_docstrings does both in one pass.

When removing a string that is a block's only statement (e.g. def f(): """doc"""), a pass is inserted so the result stays valid Python.

Installation

Install with uv:

uv add pycommremover

Usage

import pycommremover

src = '''
def greet(name):
    """Return a greeting."""          # docstring: kept by remove_comments
    msg = f"Hi, {name}"  # build it   # line comment: removed
    """a block comment"""             # block comment: removed
    return msg
'''

# Line + block comments gone, docstring kept:
print(pycommremover.remove_comments(src))

# Only the docstring gone:
print(pycommremover.remove_docstrings(src))

# Everything gone:
print(pycommremover.remove_comments_and_docstrings(src))

The input must be syntactically valid Python.

Development

The dev dependency group (installed automatically by uv run) provides pytest:

uv run pytest