BUG: Incorrect Encoding Detection as Big5

[dxdc]

OS/Arch

macOS

Python version

python 3.11

cChardet version

2.1.18

What is the problem?

cChardet is incorrectly detecting the encoding of this file as Big5. I'm not sure if the issue should be posted here or elsewhere.

import cchardet as chardet

with open('abc_1.csv', 'rb') as f:
    result = chardet.detect(f.read())
print(result)

# result:
{'encoding': 'BIG5', 'confidence': 0.9900000095367432}

Expected behavior

• The correct encoding should be one of the Windows encodings (Windows-1250, 1251, or 1252).

Actual behavior

• It is detecting the encoding of this file as Big5.

Steps to reproduce the behavior

• Use cChardet to detect the encoding of the provided CSV file.
• The encoding is incorrectly detected as Big5.

** Attached file

[wbarnha]

Thanks for providing a reference file! I see that the main project has been updated recently, so I'm going to pull in changes and see if there's any luck with the original maintainer's fixes.

[dxdc]

Thanks for the checking this out so quickly! It was difficult to create that reference file, but hopefully it gives some insights as to where the problem came from.

[dxdc]

hi @wbarnha ... did you have any luck with this one?