binaryornot 0.4.4 crashes when used with chardet 7.0.0. The new version of chardet can return {'encoding': None, 'confidence': 0.99} — high confidence, but no encoding. The condition in is_binary_string does not guard against a None encoding, so it passes None directly to .decode(), causing a crash.

Traceback

File ".../binaryornot/helpers.py", line 103, in is_binary_string
    bytes_to_check.decode(encoding=detected_encoding['encoding'])
TypeError: decode() argument 'encoding' must be str, not None

During handling of the above exception, another exception occurred:

File ".../binaryornot/helpers.py", line 106, in is_binary_string
    unicode(bytes_to_check, encoding=detected_encoding['encoding'])
NameError: name 'unicode' is not defined

Root Cause

In helpers.py, the guard condition before decoding is:

if (detected_encoding['confidence'] > 0.9 and
        detected_encoding['encoding'] != 'ascii'):

With chardet 7.0.0, detect() can return:

{'encoding': None, 'confidence': 0.99, 'language': ''}

Since None != 'ascii' evaluates to True, the code enters the if block and calls:

bytes_to_check.decode(encoding=None)  # TypeError

The except TypeError block then falls back to:

unicode(bytes_to_check, encoding=None)  # NameError: Python 2 only

So the error handler itself crashes instead of recovering gracefully.

Suggested Fix

Add detected_encoding['encoding'] is not None to the guard condition:

if (detected_encoding['confidence'] > 0.9 and
        detected_encoding['encoding'] is not None and
        detected_encoding['encoding'] != 'ascii'):

This is a minimal, surgical fix that preserves the existing logic while handling the new chardet behavior correctly.

Environment

• binaryornot: 0.4.4
• chardet: 7.0.0
• Python: 3.12.3