feat(linker): optimize text modification by processing only changed segments

[nsantacruz]

This pull request refactors how text modifications are tracked and processed in sefaria/tracker.py. The main change is to only process and log segments of text that have actually changed, rather than re-processing the entire text every time. This should improve efficiency and accuracy in change tracking and downstream processing.

Key changes:

Segment-level change tracking:

• Added a new helper function _post_modify_changed_segments that recursively compares the old and new text, and only calls post_modify_text for segments that have changed. This ensures that only modified segments are processed and logged, rather than the entire text.

• Updated modify_text to use _post_modify_changed_segments instead of calling post_modify_text directly, and adjusted how the count_after parameter is handled to avoid redundant counting and indexing.

[Copilot]

Pull request overview

This pull request refactors text modification tracking in sefaria/tracker.py to optimize performance by processing only changed segments rather than the entire text on each modification. The implementation introduces a new recursive helper function _post_modify_changed_segments that compares old and new text structures and selectively calls post_modify_text for segments that have changed.

Changes:

• Added _post_modify_changed_segments function to recursively compare and process only changed segments
• Modified modify_text to use the new segment-level tracking instead of processing the entire text
• Adjusted count_after parameter handling to prevent redundant counting during segment iteration

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[mergify]

🧪 CI Insights

Here's what we observed from your CI run for 55e96b5.

❌ Job Failures

Pipeline	Job	Health on master	Retries	🔍 CI Insights	📄 Logs
Continuous	Continuous Testing: PyTest		0	View	View

[yonadavGit]

the "kwargs all the way down" style here is a bit confusing, especially now that the counting logic has been modified. Maybe worthwhile, while were at it, taking out the the counting-related vars out of kwargs land so it's a bit clearer what's going on.. (concretely, trying to make these vars explicit as much as possible)