How manusaviour1506/quantumliverpred works
This is a research web platform for CT-slice liver-tumour modelling and comparison, not a clinical diagnostic system. The main workflow is a user uploading a slice, the backend decoding and screening it, concurrent model-family inference, and live comparison results and history. Separate branches cover labelled-slice benchmarking, whole-volume segmentation, and registry/research and system-monitoring capabilities. Edges reflect the excerpts, SOURCE INDEX references, or README; several route and persistence details are documented but their implementations were not sampled.
Main entry points: manager.py Comparison runs · patch_pipeline.py Patch preparation · registry.py Model registry · compare.py Comparison engine
Components 25
Web application
- Web client
frontend/src/App.tsx - API client
frontend/src/lib/api.ts - Session auth
frontend/src/lib/auth.tsx - Research workspace
frontend/src/pages/Research.tsx
API and services
- API application
backend/app/main.py - Comparison runs
backend/app/execution/manager.py - Benchmark service
backend/app/benchmarking/service.py - File storage
backend/app/services/storage.py - Application database
backend/app/db/database.py - System monitoring
backend/app/api/v1/system.py
CT inference
- Slice inference
src/inference/pipeline.py - Format decoding
src/preprocessing/formats.py - CT canonicalization
src/preprocessing/ct.py - Image quality checks
src/quality_control/image_qc.py - Comparison engine
src/v10/engine/compare.py - Patch preparation
src/v10/engine/patch_pipeline.py
Model families
- Model registry
src/inference/registry.py - Classical models
src/classical/nn.py - Quantum models
src/quantum/models.py - State-vector simulator
src/quantum/simulator.py - Hybrid models
src/hybrid/hybrid_model.py
Evaluation and analysis
- Labelled benchmarks
src/v10/engine/benchmark.py - Volume analysis
src/volume/pipeline.py - Segmentation metrics
src/segmentation/metrics.py
Other components
- Researcher
Connections 28
- Researcher → Web client: uses
README.md - Web client → API client: requests data
- Web client → Session auth: checks session
frontend/src/App.tsx - Session auth → API client: uses
frontend/src/lib/auth.tsx - API application → Comparison runs: initializes
backend/app/main.py - Comparison runs → Comparison engine: runs families
backend/app/execution/manager.py - Comparison runs → Application database: persists runs
- Comparison runs → File storage: stores uploads
- API application → Benchmark service: initializes
backend/app/main.py - Benchmark service → Labelled benchmarks: starts evaluation
backend/app/benchmarking/service.py - Benchmark service → Application database: stores results
backend/app/benchmarking/service.py - Slice inference → Format decoding: decodes input
src/inference/pipeline.py - Slice inference → CT canonicalization: canonicalizes image
src/inference/pipeline.py - Slice inference → Image quality checks: checks quality
src/inference/pipeline.py - Comparison engine → Patch preparation: prepares patches
src/v10/engine/compare.py - Patch preparation → Format decoding: uses decoder
src/v10/engine/patch_pipeline.py - Patch preparation → CT canonicalization: uses preprocessing
src/v10/engine/patch_pipeline.py - Patch preparation → Image quality checks: checks quality
src/v10/engine/patch_pipeline.py - Comparison engine → Model registry: selects models
src/v10/engine/compare.py - Model registry → Classical models: loads predictors
src/inference/registry.py - Model registry → Quantum models: loads predictors
src/inference/registry.py - Model registry → Hybrid models: loads predictors
src/inference/registry.py - Hybrid models → Quantum models: uses models
src/hybrid/hybrid_model.py - Quantum models → State-vector simulator: simulates circuits
src/quantum/models.py - Labelled benchmarks → Comparison engine: uses live pipeline
backend/app/benchmarking/service.py - Volume analysis → Segmentation metrics: postprocesses masks
src/volume/pipeline.py - Researcher → Research workspace: explores research
README.md - Researcher → System monitoring: monitors runtime
README.md