Skip to content

Current-State.md

Latest commit

 

History

History
120 lines (93 loc) · 6.25 KB

Current-State.md

File metadata and controls

120 lines (93 loc) · 6.25 KB

Grain Current State Assessment

Audited on 2026-04-01 against local branch main.

┌──────────────────────────────────────────────────────────────┐
│  grain · current state · 2026-04-01                         │
│  ──────────────────────────────────────────────────────────  │
│  ✓  scan → ocr → parse → save                              │
│  ✓  receipt list + detail view                             │
│  ✓  analytics: category · brand · merchant                 │
│  ✓  product index + brand catalog                          │
│  ✓  ci: build + test on every pr                           │
│  ──────────────────────────────────────────────────────────  │
│  ✗  receipt image not persisted                            │
│  ✗  edit from scan proof (todo button)                     │
│  ✗  manual entry fallback (todo button)                    │
│  ○  bank transaction import (model exists, no ui)          │
│  ○  user-facing error states (silent print)                │
└──────────────────────────────────────────────────────────────┘

Executive Snapshot

Grain has a healthy proof-of-concept core: scan receipts, parse text, store receipts, browse details, and view analytics.

The main project risk is not architecture. The main risk is product completeness and scope discipline for MVP.

Confirmed Current State

Area Current Status Notes
Platform and stack iOS 17+, SwiftUI, SwiftData, Vision, Swift Charts Aligned with ADRs
Dependencies Apple frameworks only Aligned with ADR-0003
Storage strategy Local only Aligned with ADR-0005
Core flow Scan -> OCR -> parse -> save -> list/detail Working at POC quality
Analytics Receipt/category/merchant aggregations and charts Working
Product index Product and brand indexing from receipt items Working
CI workflow Build + test workflow exists Present in .github/workflows/build.yml

Gaps and Product Debt

UX and flow gaps

  • Receipt list plus button is still a TODO (manual receipt entry not wired).
  • Scan proof Edit button is still a TODO.
  • Settings routes are placeholders (Export, Import, Tax Categories, Deduction Rules).

Data and feature gaps

  • Receipt image persistence is not wired to Receipt.imageData.
  • BankTransaction model exists but has no import, matching, or UI flow.
  • SpendingAnalytics is persisted as a model even though it behaves like derived data.

Reliability and quality gaps

  • OCR parser is still regex-based and fragile on real receipt variety.
  • Service/view errors still rely on print-based handling in key paths.
  • Unit tests exist for model coverage, but service/parser regression coverage is still limited.

Current State vs Ideal MVP

Capability Current Ideal MVP Delta
Capture and save receipt Works Works reliably with image attachment Persist imageData and add validation
Parse receipt content Works for basic receipts Robust for grocery, pharmacy, restaurant variants Add parser test corpus and parser improvements
Review and edit scanned receipt Partial User can fully correct parsed output Wire Edit flow from scan proof
Manual receipt entry Missing Available as fallback Build minimal manual entry screen
Expense analytics Works Works with trustworthy totals and period filters Add validation tests and edge-case handling
Error UX Weak User-facing recoverable errors Replace silent catches with alert states
CI confidence Moderate Green build and test on every PR Keep workflow healthy and enforce checks

PM Assessment

What is healthy

  • Architecture choices are coherent and intentionally constrained.
  • Codebase layout is understandable and suitable for iteration.
  • Delivery pace is high with active issue and PR throughput.

What needs immediate management attention

  • Backlog contains out-of-scope or conflicting items for current local-only MVP.
  • Several old issues are now completed by merged PRs and should be closed.
  • MVP definition needs explicit acceptance criteria so feature work does not sprawl.

Priority Workstream Recommendation

  ╔══════════════════════════════════════════╗
  ║  MVP PRIORITY LADDER                     ║
  ╠══════════════════════════════════════════╣
  ║  1  backlog hygiene     close stale issues ║
  ║  2  manual entry        biggest ux gap   ║
  ║  3  scan proof edit     wire todo button ║
  ║  4  image persistence   wire imageData   ║
  ╠══════════════════════════════════════════╣
  ║  5  parser reliability  test corpus      ║
  ║  6  error ux            replace print()  ║
  ╠══════════════════════════════════════════╣
  ║  ✗  cloud sync          deferred (adr-0005) ║
  ║  ✗  backend api         deferred         ║
  ╚══════════════════════════════════════════╝
  1. Stabilize backlog hygiene:
  • Close completed and out-of-scope issues.
  • Keep only active MVP issues in the next milestone lane.
  1. Close MVP usability gaps:
  • Manual entry fallback.
  • Scan proof edit flow.
  • Receipt image persistence.
  1. Improve reliability for real testing:
  • Parser regression test set.
  • User-facing error states.
  1. Delay non-MVP platform expansion:
  • Cloud sync, auth, backend API, and server-side OCR stay deferred unless ADR scope changes.