Skip to content

suggestion: enable phrase by target length instead of wordcount #7

Description

@bigpick

info

e.g to match things on such cases like https://contest-2023.korelogic.com/password-info.html

Tokenize the resulting short lines into shorter lines or phrases, typically by grabbing N characters and then moving forward until a word boundary is reached. N was ~12+ for English, lower for multibyte languages (because each character might be 2-3+ bytes long).

Activity

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

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions