Open
Description
Overview
The current profane-words repository provides a valuable list of profane terms, but it lacks the nuance to differentiate between levels of intensity and specific categories. This enhancement proposes adding two key features to enrich the dataset and allow for more flexible filtering:
- Categorization – Grouping words into relevant categories (e.g., bodily, sexual, religious, insult).
- Intensity Grading – Assigning an intensity level to each word, ranging from 1 (mild) to 5 (extremely offensive).
Proposed Data Structure
[
{
"word": "crap",
"categories": ["bodily"],
"intensity": 1
},
{
"word": "cocksucker",
"categories": ["sexual", "insult"],
"intensity": 5
},
{
"word": "damn",
"categories": ["religious"],
"intensity": 2
},
{
"word": "asshole",
"categories": ["bodily", "insult"],
"intensity": 4
}
]Usage Example
Developers could then filter words based on intensity or categories for customized moderation or content filtering:
// Filter words with intensity greater than 2
const intenseWords = profane_words.filter(pw => pw.intensity > 2);
// Filter words from the 'sexual' category
const sexualWords = profane_words.filter(pw => pw.categories.includes('sexual'));Benefits
- Customizable Filtering: Allows for more nuanced moderation levels (e.g., filtering out only the most offensive terms).
- Context-Aware Applications: Enables applications to better handle offensive content based on category relevance.
- Enhanced Clarity: Improves the readability and structure of the dataset for developers.
Implementation Considerations
- Consensus on Intensity Ratings: Establish clear criteria for assigning intensity levels.
- Backward Compatibility: Maintain the existing list format but enhance it with optional fields.
- Documentation: Update README with examples and usage instructions.
Conclusion
This enhancement will significantly improve the flexibility and usefulness of the profane-words repository. It provides more control for developers while promoting a standardized way to evaluate the severity of offensive language.
I welcome feedback on this proposal and am happy to assist with the initial implementation if approved.
Metadata
Metadata
Assignees
Labels
No labels