Skip to content

implementation_plan.md

Latest commit

 

History

History
43 lines (36 loc) · 2.15 KB

implementation_plan.md

File metadata and controls

43 lines (36 loc) · 2.15 KB
Raw

SHADOW KYC - Implementation Plan (Spec-1 Alignment)

Goal

Align the current prototype with SPEC-1-Shadow-Real-Time-KYC-Integrity-Validator, focusing on Role Separation (User/Tenant), Session Gateway, and Database Integration.

Current State

  • Frontend: Single Dashboard combining Camera + Analysis.
  • Backend: Single WebSocket endpoint (/ws/live) processing frames and returning results to the same socket. No persistence.
  • Layers: 5 Implemented (Quality, Artifact, Liveness, Temporal, Deepfake). Spec requires 7 specific layers.

Phase 1: Session Gateway & Backend logic (Priority)

Enable the "User streams, Tenant watches" workflow.

  1. Session Manager (app/core/session_manager.py):
    • Manage active sessions in memory.
    • Support two roles per session: CLIENT (sends video) and TENANT (receives analytics).
    • Generate 6-digit session codes.
  2. WebSocket Endpoint Update (app/main.py):
    • Update /ws/live/{client_type}/{session_id}.
    • Route frames from CLIENT to Orchestrator.
    • Broadcast results to TENANT.

Phase 2: Frontend Roles & Routing

  1. Dependencies: Install react-router-dom.
  2. Routes:
    • / -> User Portal (Input Session Code or Upload).
    • /session/:code -> Live Client (Camera only, simple UI status).
    • /tenant -> Tenant Portal (Login/Create Session).
    • /tenant/monitor/:code -> Integrity Dashboard (The current complex UI).
  3. Refactor:
    • Extract VideoInput logic to user/LiveClient.jsx.
    • Extract Dashboard logic to tenant/Monitor.jsx.

Phase 3: Database Integration (PostgreSQL/SQLite)

  1. Database Module: Setup app/core/database.py using SQLAlchemy.
  2. Models: Session, Evidence, Score.
  3. Persistence: Save final reports to DB in orchestrator.finalize_session.

Phase 4: Algorithm Refinement (Layers 1-7)

  1. Layer Extension: Implement placeholders for L6 (Environment) and L5 (Lip-Sync).
  2. Meta Fusion: Refactor RiskEngine to MetaFusionEngine with specific "Explainable AI" output tags.

Next Step: Phase 1 (Backend Session Manager) & Phase 2 (Frontend Routing) in parallel.