Sangeetha-Grantha

Metadata	Value
Status	Active
Version	1.1.0
Last Updated	2026-09-10
Author	Sangeetha Grantha Team
Document Type	Evidence record

TextBlocker Parsing Fix Implementation

---

[!NOTE] Historical evidence: results, counts, commands, and observations below belong to the original work described here. The editorial update date is not a new test or corpus verification. For present behavior, use current ingestion guide.

1. Purpose

Fix issues in TextBlocker regex logic where section headers with suffixes (e.g., (madhyama kAla sAhityam)) or inline headers (e.g., P cakkani...) were not being correctly identified, causing headers to leak into lyric content or sections to be missed entirely.

2. Changes

• Refactored TextBlocker.kt:
  • Updated detectSectionHeader to support:
    • Optional suffixes like sahityam (e.g., Madhyama Kala Sahityam).
    • caranam spelling (starting with ‘c’).
    • Surrounding punctuation removal (parentheses, brackets).
    • Single-letter abbreviations (P, A, C) when at the start of a line.
    • Devanagari headers (पल्लवि, अनुपल्लवि, etc.).
  • Robustness: Added automatic unescaping of literal \n strings before line splitting.
• Enhanced ImportService.kt: Added sanitization logic to unescape \n literals during the review and section-mapping flow, preventing data corruption from reaching the structural tables.
• Verification: Validated against multiple blog formats (Thyagaraja Vaibhavam, Syama Krishna Vaibhavam) covering inline headers and complex transliteration. Corrected existing data for krithi c64243b2 using a focused fix script.

3. Commit Reference

This document serves as the Ref for the commit.

4. Progress

• Identify regex flaws.
• Implement robust header detection.
• Verify against “curve ball” URLs.

---

Documentation home · Feature status