How degenerateuser/mplads-review works
This is a local-first MPLADS anomaly-review decision-support prototype. Its main workflow is to load provenance-labeled works, detect review signals, decompose risk and evidence, present role-oriented queues and comparisons, then save a human review decision. The README documents the product capabilities and JSON-backed demo architecture; source-index references and excerpts support the shown internal links. Several detector and route implementations were not sampled, so their detailed call paths are omitted.
Main entry points: analysis.py Analysis and evidence · models.py Domain models · App.tsx Dashboard and review · api.ts API client
Components 17
Review experience
- Dashboard and review
frontend/src/App.tsx - Evidence comparison
frontend/src/components/EvidencePanel.tsx - API client
frontend/src/api.ts - Demo fallback data
frontend/src/demo.ts - Session context
frontend/src/context/AuthContext.tsx
API and identity
- FastAPI routes
backend/main.py - Authentication
backend/auth.py
Anomaly analysis
- Analysis and evidence
backend/analysis.py - Duplicate detection
backend/detectors/duplicate.py - Unit-cost detection
backend/detectors/cost.py - Stalled-work detection
backend/detectors/stalls.py - Photo review
backend/detectors/photos.py - Risk scoring
backend/scoring.py
Records and persistence
- Domain models
backend/models.py - Synthetic work generator
data/generator.py - JSON and review storage
backend/repository.py
Other components
- Human reviewer
Connections 24
- Human reviewer → Dashboard and review: selects work
README.md - Dashboard and review → API client: requests summary
frontend/src/App.tsx - API client → FastAPI routes: calls API
README.md - API client → Demo fallback data: uses fallback
frontend/src/api.ts - Dashboard and review → Evidence comparison: shows evidence
frontend/src/App.tsx - Human reviewer → Dashboard and review: submits review
frontend/src/App.tsx - Dashboard and review → API client: saves decision
frontend/src/App.tsx - FastAPI routes → Analysis and evidence: serves analysis
backend/main.py - FastAPI routes → JSON and review storage: loads records
backend/main.py - FastAPI routes → Authentication: authenticates users
backend/main.py - Session context → API client: calls auth methods
frontend/src/context/AuthContext.tsx - Analysis and evidence → Duplicate detection: runs detector
backend/analysis.py - Analysis and evidence → Unit-cost detection: runs detector
backend/analysis.py - Analysis and evidence → Stalled-work detection: runs detector
backend/analysis.py - Analysis and evidence → Photo review: runs detector
backend/analysis.py - Analysis and evidence → Risk scoring: computes scores
backend/analysis.py - Analysis and evidence → Domain models: uses domain types
backend/analysis.py - Duplicate detection → Domain models: reads works
backend/detectors/duplicate.py - Unit-cost detection → Domain models: reads works
backend/detectors/cost.py - Stalled-work detection → Domain models: reads works
backend/detectors/stalls.py - Photo review → Domain models: reads photo metadata
backend/detectors/photos.py - Risk scoring → Domain models: uses score types
backend/scoring.py - Synthetic work generator → Domain models: creates works
data/generator.py - JSON and review storage → Domain models: loads and saves
backend/repository.py