feat(semantic_dedup): configurable similarity threshold and shingle size#10
Open
rambalconnicz wants to merge 1 commit intoopen-compress:mainfrom
Open
Conversation
…igurable The SemanticDedup stage previously used hardcoded values for the Jaccard similarity threshold (0.8) and shingle size (3). Different content types benefit from different dedup aggressiveness -- e.g. log-heavy output works better with a looser threshold (0.6) while code blocks need stricter matching (0.9) to avoid false merges. This adds optional constructor parameters: - similarity_threshold: Jaccard cutoff (default 0.8, unchanged) - shingle_size: n-gram width for fingerprinting (default 3) - min_block_chars: minimum block length to consider (default 50) All parameters are threaded through _run_dedup and _split_blocks. Existing behavior is preserved when no arguments are passed.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The
SemanticDedupstage uses a hardcoded Jaccard similarity threshold of 0.8 and shingle size of 3. This works well for general text, but different content types benefit from different dedup aggressiveness:This PR makes these parameters configurable via
SemanticDedupconstructor args while preserving the existing defaults.Changes
similarity_threshold,shingle_size, andmin_block_charsconstructor params toSemanticDedup_run_dedup()and_split_blocks()Usage