Sangeetha-Grantha

Metadata	Value
Status	Completed
Version	1.0.1
Last Updated	2026-02-02
Author	Sangeetha Grantha Team
Related Tracks	TRACK-001 (Bulk Import), TRACK-008 (Entity Resolution)

---

TRACK-031: Composer Deduplication – Avoid Duplicate Composer Names

1. Goal

Avoid creating duplicate composer records when the same person is referred to by different names (e.g. “Dikshitar” vs “Muthuswami Dikshitar”, “Thyagaraja” vs “Tyagaraja”). Ensure a single canonical composer per person and map aliases to it during import and reference-data entry.

2. Problem Statement

• Observed: Reference data lists show “Dikshitar” (Unknown Place • ? - ?) and “Muthuswami Dikshitar” (1775 - 1835) as separate composers; they are the same person.
• Cause: Scraped or manual input uses short forms (“Dikshitar”); entity resolution may suggest candidates but review flow can still create a new composer instead of mapping to the existing canonical one. No alias/canonical rules or merge flow.
• Impact: Duplicate composers fragment krithi attribution, complicate search, and clutter reference data.

3. Scope (Analysis)

• In scope: Composers (primary); patterns may extend to Ragas/Talas alias handling.
• Out of scope: Merging existing duplicate records in DB (can be a follow-up track).

4. Implementation Plan (To Be Refined)

4.1 Canonical Names and Aliases

• Define canonical composer names (e.g. “Muthuswami Dikshitar”, “Tyagaraja”, “Syama Sastri”) and a small alias table or config (e.g. “Dikshitar” → “Muthuswami Dikshitar”).
• Options: DB table composer_aliases (alias_normalized, composer_id) or config-driven map; prefer DB for maintainability.

4.2 Entity Resolution

• In EntityResolutionService / NameNormalizationService: before fuzzy match, resolve known aliases to canonical composer and match by canonical id.
• When resolution returns HIGH confidence for an alias → return the canonical composer as the single candidate (no “create new” suggestion for the alias).

4.3 Import Review Flow

• On approve, if user selects a composer from resolution candidates, use that composer id.
• If user does not select (or overrides with a new name): resolve alias → canonical before create; if alias maps to existing canonical, use that composer id and do not create a new composer.
• findOrCreate: When creating a composer by name, normalize and check alias map first; if alias exists, return existing canonical composer.

4.4 Reference Data UI

• Composer create/edit: warn or block when the normalized name matches an existing composer or a known alias of another composer.
• Optional: “Merge composer” action to merge a duplicate into a canonical one (data migration). (Deferred to future track)

4.5 Default Language for Composers

• Majority of Dikshitar compositions are in Sanskrit; consider defaulting primary language for krithis when composer is identified as Dikshitar (or from composer metadata) to SA when not specified. (Can be a small follow-up.)

5. References

• TRACK-008 – Entity Resolution Hardening (composer/raga/tala normalization).
• Bulk Import Strategy – Lyric structure and entity mapping.
• Source example: Guru Guha – Bala Kuchambike (composer cited as “Dikshitar” in context of Muthuswami Dikshitar).

6. Progress Log

• 2026-01-30: Track created; analysis and implementation plan drafted.
• 2026-01-30: Implemented composer_aliases migration (22__composer_aliases.sql), ComposerAliasesTable, ComposerAliasRepository, seed 03_composer_aliases.sql; wired alias resolution into EntityResolutionService (alias map in composerMap) and ComposerRepository.findOrCreate (resolve alias before create). Reference data UI warning/merge deferred.
• 2026-02-01: Fixed composer alias lookup to avoid implicit join without FK, preventing bulk approve failures during import review.
• 2026-02-02: Track completed. Alias mechanism verified.