How koraygokceler/rf_cavity_neural_operator works
RF Cavity Neural Operator generates FEM-labelled cavity data, trains a field-prediction model, and serves predictions and visualization through a web interface. The graph preserves the documented data-to-training workflow and the main service, QoI, and display layers. Edges are limited to README statements and sampled source references; several implementation details, including frontend requests and job orchestration, were not fully sampled.
Main entry points: model.py Model service · api.py Web API · field_dataset.py Field dataset · dataset_generator_3d.py Dataset generator
Components 23
Data generation
- Cavity shapes
src/data_gen/cavity_shapes.py - Dataset generator
src/data_gen/dataset_generator_3d.py - Field labels
src/data_gen/field_labels.py - High-order FEM
src/data_gen/highorder.py - Field dataset
src/data/field_dataset.py - Mesh features
src/data/mesh_features.py
Model and training
- Training entry
train.py - Training module
src/training/lightning_module.py - Field operator
src/models/eigenspace_operator_field.py - H(curl) algebra
src/models/hcurl.py - Model layers
src/models/layers.py
Prediction service
- Web API
src/service/api.py - Geometry service
src/service/geometry.py - Model service
src/service/model.py - Background jobs
src/service/jobs.py
Web interface
- Web application
web/src/App.tsx - Workspace
web/src/components/Workspace.tsx - API client
web/src/api.ts - Field view
web/src/components/useFieldView.ts - 3D viewer
web/src/components/Viewer3D.tsx
QoI and visualization
- Cavity figures
src/qoi/figures.py - Field display
src/viz/hcurl_display.py
Other components
- Engineer
Connections 27
- Engineer → Dataset generator: generates data
README.md - Dataset generator → Cavity shapes: draws geometry
src/data_gen/dataset_generator_3d.py - Dataset generator → Field labels: creates labels
src/data_gen/dataset_generator_3d.py - Field labels → High-order FEM: solves modes
src/data_gen/field_labels.py - Field labels → Cavity figures: computes QoI
src/data_gen/field_labels.py - Field dataset → High-order FEM: builds operators
src/data/field_dataset.py - Field dataset → Mesh features: builds features
src/data/field_dataset.py - Training entry → Field dataset: loads samples
train.py - Training entry → Training module: starts training
train.py - Training module → Field operator: trains model
src/training/lightning_module.py - Field operator → H(curl) algebra: uses algebra
src/models/eigenspace_operator_field.py - Field operator → Model layers: uses layers
src/models/eigenspace_operator_field.py - Engineer → Web API: requests prediction
src/service/api.py - Web API → Geometry service: checks geometry
src/service/api.py - Web API → Model service: requests prediction
src/service/api.py - Model service → Geometry service: loads mesh
src/service/model.py - Model service → Field dataset: builds model item
src/service/model.py - Model service → High-order FEM: builds operators
src/service/model.py - Model service → Cavity figures: computes figures
src/service/model.py - Model service → Field display: prepares fields
src/service/model.py - Web API → Background jobs: dispatches jobs
src/service/api.py - Web application → Workspace: shows workspace
web/src/App.tsx - Workspace → API client: uses client
web/src/components/Workspace.tsx - Workspace → Field view: renders field view
web/src/components/Workspace.tsx - Field view → 3D viewer: uses viewer helpers
web/src/components/useFieldView.ts - API client → Web API: calls endpoints
- 3D viewer → Field display: renders fields