Data: Validate and clean profanity dataset (duplicates & false positives)

[master-wayne7]

The current profanity dataset is sourced from multiple language files and contains ~55K+ entries.

We need to improve the quality of this dataset by:

Tasks:

• Identify and remove duplicate entries across languages
• Detect and eliminate false positives (valid words incorrectly flagged)
• Normalize formatting (case sensitivity, spacing, special characters)
• Ensure consistency across all language files

Optional:

• Add a script/tool to automate dataset validation

This will significantly improve filtering accuracy and reduce unintended masking.