This PR fixes a bug where passing a valid bytearray to from_bytes crashes the application with a TypeError. The api.from_bytes type hints explicitly support bytearray, but the underlying helper function utils.any_specified_encoding did not.

What actually happens:
When from_bytes is called with a bytearray and preemptive_behaviour=True, utils.any_specified_encoding raises a TypeError because it strictly checks if not isinstance(sequence, bytes):.

What I expected to happen:
The function should successfully process the bytearray sequence and return a CharsetMatches object, just as it does for standard bytes objects.

How to reproduce the issue:

from charset_normalizer import from_bytes

# Create a bytearray object 
my_bytearray = bytearray(b"Hello, world!")

# This will raise a TypeError on the current main branch
results = from_bytes(my_bytearray)

Changes Made:

• Updated utils.py around line 260 to check for isinstance(sequence, (bytes, bytearray)) instead of strictly bytes.
• This fix ensures backward compatibility while aligning the runtime behavior with the documented type hints.

Checklist:

• I have read the CONTRIBUTING.md document.
• I have verified that this does not break backward compatibility.
• I have successfully run nox -s test locally.
• I have successfully run nox -s lint locally.
• I have successfully run nox -s coverage locally.