Fix performance regression and simplify tag handling #23
Conversation
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
Group coalescing patterns by their expected first character ({, <, [, backtick)
so we only check relevant patterns instead of all 891 pattern groups.
Result: 19% faster (339ms vs 419ms in profile), now 54% faster than v0.6.0
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Move tag-specific pattern categorization logic from text_wrapping.py to tag_handling.py. Add get_tag_coalescing_patterns_by_start() that returns patterns already grouped by expected start character. This improves separation of concerns: tag_handling knows which patterns match which delimiters, while text_wrapping just uses the grouped data. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Replace the complex coalescing-based word splitting with a simpler single-pass regex extraction approach: 1. Define atomic patterns (code spans, links, tags) using frozen AtomicPattern dataclasses for clarity and maintainability 2. Use a single compiled regex that matches all atomic constructs in one pass, extracting them as placeholders before word splitting 3. Split on whitespace, then restore original constructs This approach is simpler (fewer lines), faster (one regex pass vs per-word pattern matching), and keeps Markdown links as atomic units which improves wrapping behavior. Performance: 57% faster than v0.6.0 (96ms vs 225ms median). Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Remove `--tags` CLI option and `TagWrapping` enum (atomic mode is now always-on) - Simplify `AtomicPattern` dataclass to just `name` and `pattern` fields - Define delimiter constants directly as module-level variables - Remove dead code: coalescing patterns, placeholder extraction from tag_handling - Remove `tags` parameter from all API functions (reformat_text, fill_markdown, etc.) - Move completed atomic-tag-wrapping specs to done/ The atomic tag wrapping behavior is now the only mode - template tags, code spans, and markdown links are always kept together as indivisible tokens during line wrapping. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Remove module-level delimiter constants (JINJA_TAG_OPEN, etc.) - Add required `open_delim`, `close_delim`, `open_re`, `close_re` fields to AtomicPattern - Update tag_handling.py to use pattern properties directly (e.g., SINGLE_JINJA_TAG.open_delim) - Remove unnecessary `or ""` guards since fields are now always strings Each pattern now contains all its information in one place, making the code cleaner and easier to maintain. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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
make benchmark,make profile) for performance testing--tagsCLI option (atomic mode is now always-on)AtomicPatterndataclassAtomicPatterndataclassPerformance
Before fix: ~2900ms (29x regression)
After fix: ~100ms (2.2x faster than v0.6.0 baseline of ~220ms)
Key Changes
Performance fix: Moved
preprocess_tag_block_spacingbefore Markdown parsing to avoid repeated traversalsBenchmarking tools: New
devtools/benchmark.pywith--compare,--profile, and--semanticoptionsAPI simplification: Removed
--tagsoption andTagWrappingenum - atomic tag handling is now the default and only modeCode consolidation:
atomic_patterns.pymodule withAtomicPatterndataclassTest plan
🤖 Generated with Claude Code