Skip to content

[BUG] BERTUnfactoredDisambiguator.pretrained() clips text when special character � exists #159

Description

@ahmadabousetta

Describe the bug
BERTUnfactoredDisambiguator.pretrained() clips last token of text when text contains the special character �.
MLEDisambiguator.pretrained() works fine and doesn't clip tokens.

To Reproduce

from camel_tools.disambig.mle import MLEDisambiguator
from camel_tools.disambig.bert import BERTUnfactoredDisambiguator

mle = MLEDisambiguator.pretrained()
brt = BERTUnfactoredDisambiguator.pretrained()

mle.disambiguate("أعضاء مجلس � الإدارة المحترمون يتفضلون".split())
brt.disambiguate("أعضاء مجلس � الإدارة المحترمون يتفضلون".split())

Expected behavior
To provide disambiguation for all tokens.

Screenshots

Image

Desktop (please complete the following information):

  • WSL2 under Windows 11 home.
  • WSL version: 2.5.9.0
  • Kernel version: 6.6.87.2-1
  • WSLg version: 1.0.66
  • MSRDC version: 1.2.6074
  • Direct3D version: 1.611.1-81528511
  • DXCore version: 10.0.26100.1-240331-1435.ge-release
  • Windows version: 10.0.26100.4349
  • Python version: 3.12
  • CAMeL Tools version as well as installation source (pip, conda, source). camel_tools.version '1.5.6' using pip

Additional context
I understand it's not an expected character in Arabic text. But it happens occasionally in automated workflows. Thank you.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions