Sangeetha-Grantha

Metadata	Value
Status	Completed
Version	1.0.0
Last Updated	2026-03-20
Author	Sangeetha Grantha Team

Track: Strip Residual Numbers from Numbered Section Labels

ID: TRACK-101 Status: Not Started Owner: Sangita Grantha Architect Created: 2026-03-20 Updated: 2026-03-20 Parent: TRACK-093 (Trinity Krithi Import)

Goal

Fix the StructureParser so that when numbered section headers like caraNam 1, svara sAhitya 2 are matched, the trailing number is stripped from the section body text. Currently the number bleeds into the lyric content as a leading character.

Context

• Investigation Report: database/for_import/EXTRACTION-INVESTIGATION-REPORT.md
• Primary File(s): tools/krithi-extract-enrich-worker/src/structure_parser.py
• Severity: Secondary
• Affected Blogs: Primarily Syama Sastri (numbered caraNam 1/2/3) and any blog with numbered sections
• Current Behavior: When _detect_section_header() matches caraNam 1, the regex pattern.sub("", line, count=1) at line 399 removes the matched section keyword but leaves the trailing number. The remainder becomes "1 \n mANikya mayamaiyunna...".
• Expected Behavior: The remainder should be "mANikya mayamaiyunna..." with no leading number.

Analysis

Verified Code Path

1. _detect_section_header() (line 396) iterates SECTION_HEADER_PATTERNS
2. The CHARANAM pattern (line 146-149): r"^\s*[\-–—•*()=\[\]]*\s*(?:(?:ch|c)ara?n(?:\.\s*am|am)|caraṇam)(?:\b|:|\.|\-|\)|]|=|$)"

3. This pattern matches caraNam but NOT the trailing ` 1 — the (?:\b

   …

   $)` anchor group matches at the word boundary after “caraNam”

4. Line 399: remainder = pattern.sub("", line, count=1).strip() removes caraNam but " 1" stays
5. Line 400: remainder = re.sub(r"^[:\-)\]\.\s]+", "", remainder).strip() strips leading punctuation/whitespace but NOT digits
6. The remainder "1" is then prepended to the section text via _HeaderMatch.remainder

Fix Location

The fix belongs in _detect_section_header() at line 400. After stripping punctuation/whitespace, also strip a leading number (with optional whitespace) that represents the section occurrence counter.

# Current (line 400):
remainder = re.sub(r"^[:\-)\]\.\s]+", "", remainder).strip()

# Fixed:
remainder = re.sub(r"^[:\-)\]\.\s]+", "", remainder).strip()
remainder = re.sub(r"^\d+\s*", "", remainder).strip()

Edge Case Consideration

Need to ensure this doesn’t strip leading digits that are actually part of lyric text. In Carnatic compositions, lyric lines don’t start with bare numbers — numbered section markers are the only source of leading digits after a header match. Safe to strip.

Implementation Plan

Phase 1: Fix

• In _detect_section_header() (line 400), add re.sub(r"^\d+\s*", "", remainder) after the existing punctuation strip
• Alternatively, extend the existing regex at line 400 to include digits: r"^[:\-)\]\.\s\d]+"

Phase 2: Testing & Validation

• Unit test: input "caraNam 1\nmANikya mayamaiyunna" → section text should be "mANikya mayamaiyunna" with no leading "1"
• Unit test: input "svara sAhitya 2\nkunda radanA" → section text "kunda radanA"
• Unit test: input "caraNam\nmANikya" (no number) → still works correctly
• Regression: verify existing passing extractions still work
• Integration test against Syama Sastri blog HTML (has numbered caraNam 1/2/3)

Design Considerations

Simple one-line fix. The trailing-number strip should happen after the existing punctuation strip to handle cases like caraNam - 1 or caraNam: 2.

Acceptance Criteria

• No leading digits in section body text when section header has a number suffix
• Sections without numbers are unaffected
• All existing tests pass
• Validated against real HTML from all 3 blog sources

Dependencies

• Depends on: TRACK-100 (multi-pass parsing must work first so we can validate across all variants)
• Blocks: None

Files to Modify

• tools/krithi-extract-enrich-worker/src/structure_parser.py — _detect_section_header() method, line ~400
• tools/krithi-extract-enrich-worker/tests/test_structure_parser.py — add numbered-label test cases