| Metadata | Value |
|---|---|
| Status | Active |
| Version | 1.2.0 |
| Last Updated | 2026-09-10 |
| Author | Sangeetha Grantha Team |
| Document Type | Navigation |
This is the current capability map for Sangeetha Grantha, reviewed against repository code on 2026-09-10. Use it to distinguish working behavior, operational prerequisites, and planned experiences. Track reports record past verification; they do not establish the state of a running deployment.
| Capability | Implemented behavior | Boundary / next work |
|---|---|---|
| Rasika shell | Home, Explore, Library, Settings; persistent appearance | Native-device acceptance remains open |
| Catalogue discovery | V2 discovery response; published selection and available feature content | Expanded editorial Home features remain later-release work |
| Search and browse | Explicit submitted query across compositions, ragas, composers; paging; UUID filters | Tala, deity, temple, language, and form directories are not mounted in V2 |
| Entity pages | Raga/composer detail and works lists; aliases and raga lineage where available | Displayed metadata depends on stored reference coverage |
| Lyric reader | One source variant at a time; script/source selection; stored section labels; ordered ragas | No generated transliteration, playback, or notation renderer in the Rasika reader |
| Library | Device-local bookmarks across restart | Account sync, collections, and recents remain planned |
| Accessibility | Large-text handling and reduced-motion behavior; native journey test targets | TalkBack, VoiceOver, and device runtime proof are still required |
Read the mobile requirements, Rasika UI guide, and TRACK-140 evidence.
| Capability | Implemented behavior | Boundary / next work |
|---|---|---|
| Catalogue editing | Metadata, structure, lyric variants, notation, tags, audit views | Musical form and section provenance require curator judgment |
| Search modes | Lexical, hybrid, and semantic selection in the admin list | Vector modes require a compatible active embedding profile and indexed documents |
| Bulk import | Manifest upload, batches, jobs/tasks/events, pause/resume/retry/cancel, review and finalization | Extraction success alone does not mean publication |
| Sourcing workspace | Sources/extractions, evidence/verification, quality summary screens | Coverage and audit-summary endpoints include placeholder responses |
| Curator review | Section issues and raga-resolution actions | Corpus corrections should be verified after reingestion |
| Reference data | Composer/raga/tala/deity/temple management; raga tree, aliases, identity rules | Ambiguous names are reviewed; similarity does not establish raga identity |
| Authentication | Admin-token exchange for JWT, stored role checks, JWT refresh; argon2id provisioning | Password provisioning is not an interactive password login; OAuth/OTP is deferred |
Read the curator guide, bulk-import guide, and authentication reference.
| Capability | Where to learn more |
|---|---|
| Public V1/V2 contracts, published visibility, strict query parsing | API contract |
| Embedding profiles, overview/passage indexing, hybrid rank fusion | Search guide |
| Append-only revisions and per-section source provenance | Versioned canon |
| Raga aliases, authority, match keys, resolution queue | Raga identity |
| Python extraction, canonical payloads, Kotlin persistence | Ingestion architecture |
| Flyway schema and repeatable reference seeds | Migrations |
| Backend, frontend, worker, and mobile checks | Quality guide |
The documents below explain feature design or earlier implementation work. Check their lifecycle notice and the current guides before treating a proposal as shipped behavior.
The product requirements describe intended outcomes. The Conductor registry owns task status. The documentation index provides the complete reading map.