fix: handle missing 'import' keyword in break_up_import

[frankgoldfish]

Problem

When source code contains carriage returns (`\r`) within parenthesized import statements, io.StringIO.readlines() may produce continuation lines that don't contain the import keyword. When break_up_import() is called on such a line, re.split(r'\bimport\b', line) returns only one element, and unpacking to (indentation, imports) raises:

ValueError: not enough values to unpack (expected 2, got 1)

Reproducible with:

import autoflake
s = '''
from t import (
\r  a,
    b,
)
from t import (
    a,
    b,
)
'''
autoflake.fix_code(s, remove_all_unused_imports=True)  # raises ValueError

Fix

In break_up_import(), check the number of parts returned by re.split(). If 'import' is not found in the line (i.e., it's a bare continuation line), return the line unchanged.

Tests

Added TestCarriageReturnInImport test class verifying the fix.

Fixes #326

[frankgoldfish]

Closing this PR — found that PR #328 already addresses this issue. Apologies for the duplicate.