Roadmap¶

Current development status, known TODOs from the source code, and planned features for the NassaQ platform.

Current Status¶

Implementation Overview¶

pie title Implementation Status
    "Implemented" : 12
    "Partial / Mock" : 7
    "Planned" : 6

Source Code TODOs¶

These TODOs are extracted directly from the codebase:

OCR Worker (ocr/app/services/worker.py)¶

Context: The OCR worker currently saves processed output (OCR text, metadata, confidence scores) to local files in the /ocr/documents directory. The plan is to migrate this to MongoDB for centralized, queryable storage.

Backend Server¶

Planned Features¶

High Priority¶

1. MongoDB Migration for OCR Output¶

Current state: OCR results are saved as local files (3 files per document: text, metadata JSON, confidence data). See the Processing Pipelines page for details on the current output format.

Target state: Store all OCR output in MongoDB collections for:

• Full-text search across processed documents
• Structured metadata queries
• Centralized storage accessible by all services
• Elimination of local file management

flowchart LR
    subgraph Current
        OCR1[OCR Worker] --> FS[Local Files<br/>/ocr/documents/]
    end

    subgraph Planned
        OCR2[OCR Worker] --> MDB[MongoDB<br/>OCR results collection]
        SRV[Backend Server] --> MDB
        FE[Frontend] -->|via API| SRV
    end

2. Frontend-Backend Integration¶

Connect mock data pages to existing backend endpoints. For a list of which frontend pages currently use mock data, see the Components & Flows page.

3. Document Upload UI¶

Build a document upload interface connecting to POST /api/v1/docs/upload:

• File selection with drag-and-drop
• Virtual path (folder) selection
• Upload progress indicator
• Automatic status polling after upload

Medium Priority¶

4. TanStack React Query Migration¶

Current state: API calls use raw apiFetch() in useEffect hooks with manual loading/error state management. See the Frontend API Integration page for the current state of the configured-but-unused QueryClient.

Target state: Migrate to React Query for:

• Automatic caching and deduplication
• Background refetching
• Optimistic updates
• Declarative loading/error states
• QueryClient is already configured but unused

5. Azure Service Bus¶

Current state: RabbitMQ is used for development. The backend has an abstract BaseBroker class and a stub AzureServiceBusBroker. See the Deployment page for the current Docker Compose setup.

Target state: Implement AzureServiceBusBroker for production use, providing:

• Managed service (no infrastructure to maintain)
• Built-in dead letter queues
• Topic-based routing
• Integration with Azure monitoring

6. Token Refresh Hardening¶

Current state: Refresh tokens are not rotated -- the same refresh token is reused for 7 days. See the Security & Auth page for current JWT implementation details.

Improvements needed:

• Refresh token rotation (new refresh token on each use)
• Server-side refresh token storage for revocation
• Logout endpoint that invalidates refresh tokens

7. Rate Limiting¶

Add rate limiting on sensitive endpoints:

Low Priority¶

8. Password Reset Flow¶

Implement email-based password reset:

1. User requests reset via email
2. Server sends time-limited reset token
3. User clicks link with token
4. User sets new password

Requires email service integration (e.g., Azure Communication Services).

9. Individual Permissions¶

The Individual_Permissions database table exists but is not used. Implement per-document sharing:

• Grant read/write access to specific users on specific documents
• Permission inheritance through virtual path hierarchy
• Permission management UI

10. Audit Logging¶

The Logs database table exists but is not populated. Implement:

• Authentication events (login, logout, failed attempts)
• Admin actions (user activation, role changes, deletions)
• Document operations (upload, delete, status changes)
• API access logging for compliance

11. Frontend Docker Containerization¶

The frontend currently runs outside Docker Compose. Add a containerized build:

# Planned addition to docker-compose.yml
frontend:
  build:
    context: ./frontend
    dockerfile: Dockerfile
  ports:
    - "8080:80"
  depends_on:
    - server

Known Limitations¶

Backend¶

OCR Worker¶

Frontend¶