# rita12321/nndosenet architecture

> Architecture diagram of the GitHub repository rita12321/nndosenet, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-07). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/rita12321/nndosenet
- Repository: https://github.com/rita12321/nndosenet

## Overview

nnDoseNet is a radiotherapy dose-prediction framework. Its principal workflow converts and labels DICOM/NIfTI cases, plans and preprocesses datasets, trains cross-validation models, predicts dose maps, and evaluates clinical and image metrics. Separate capabilities include DICOM dose export and gamma analysis, plus model ensembling. The graph emphasizes these documented stages and source-index-supported calls; unsampled subsystems are included where documented, with uncertain connections omitted.

Main entry points: [`nndosenet/inference/predict_from_raw_data.py`](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/inference/predict_from_raw_data.py) (Prediction runner), [`nndosenet/training/nnDoseTrainer/nnDoseTrainer.py`](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/training/nnDoseTrainer/nnDoseTrainer.py) (Dose trainer), [`nndosenet/evaluation/evaluate_predictions.py`](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/evaluation/evaluate_predictions.py) (Dose evaluation), [`nndosenet/experiment_planning/plan_and_preprocess_entrypoints.py`](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/experiment_planning/plan_and_preprocess_entrypoints.py) (Plan entrypoints)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_data["Dataset preparation"]
  node_format["Format and label"]
  node_fingerprint["Dataset fingerprint"]
  node_preprocess["Preprocessing"]
  node_resample["Resampling"]
end

subgraph group_model["Planning and training"]
  node_planning["Experiment planning"]
  node_plan_api["Plan entrypoints"]
  node_trainer["Dose trainer<br/>[nnDoseTrainer.py]"]
  node_train_run["Training runner<br/>[run_training.py]"]
  node_loss["DVH loss<br/>[DVH_loss.py]"]
  node_dataloader["Training data loading"]
end

subgraph group_prediction["Inference and evaluation"]
  node_predictor["Prediction runner"]
  node_iterator["Prediction data iterator<br/>[data_iterators.py]"]
  node_sliding["Sliding-window prediction"]
  node_export["Prediction export"]
  node_evaluation["Dose evaluation"]
  node_dicom_eval["DICOM evaluation"]
  node_gamma["Gamma analysis<br/>[gamma_analysis.py]"]
  node_ensemble["Model ensembling<br/>[ensemble.py]"]
end

subgraph group_support["Medical imaging support"]
  node_dicom_in["DICOM conversion<br/>[dicom2nifti.py]"]
  node_dicom_out["NIfTI to DICOM"]
  node_imageio["Image I/O registry"]
  node_reader["SimpleITK reader"]
  node_plans["Plans and labels<br/>[plans_handler.py]"]
end

node_user(("Researcher"))

node_user -->|"provides cases"| node_format
node_format -->|"converts DICOM"| node_dicom_in
node_user -->|"requests planning"| node_plan_api
node_plan_api -->|"runs planner"| node_planning
node_planning -->|"analyzes dataset"| node_fingerprint
node_plan_api -.->|"preprocesses cases"| node_preprocess
node_preprocess -->|"resamples images"| node_resample
node_user -->|"starts training"| node_train_run
node_train_run -->|"instantiates trainer"| node_trainer
node_trainer -->|"loads batches"| node_dataloader
node_trainer -.->|"computes DVH loss"| node_loss
node_trainer -->|"evaluates validation"| node_evaluation
node_user -->|"requests prediction"| node_predictor
node_predictor -->|"iterates inputs"| node_iterator
node_predictor -->|"predicts patches"| node_sliding
node_predictor -->|"exports predictions"| node_export
node_predictor -->|"computes metrics"| node_evaluation
node_evaluation -->|"selects reader"| node_imageio
node_imageio -->|"registers reader"| node_reader
node_predictor -->|"loads plan metadata"| node_plans
node_user -->|"requests RT analysis"| node_dicom_eval
node_dicom_eval -->|"converts dose"| node_dicom_out
node_dicom_eval -->|"runs gamma analysis"| node_gamma
node_user -.->|"combines models"| node_ensemble
node_ensemble -->|"evaluates ensemble"| node_evaluation

click node_format "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/formating_dataset/format_dataset_api.py"
click node_dicom_in "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/dicom_bridge/dicom2nifti.py"
click node_fingerprint "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/experiment_planning/dataset_fingerprint/fingerprint_extractor.py"
click node_planning "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/experiment_planning/experiment_planners/default_experiment_planner.py"
click node_plan_api "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/experiment_planning/plan_and_preprocess_entrypoints.py"
click node_preprocess "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/preprocessing/preprocessors/default_preprocessor.py"
click node_resample "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/preprocessing/resampling/default_resampling.py"
click node_trainer "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/training/nnDoseTrainer/nnDoseTrainer.py"
click node_train_run "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/run/run_training.py"
click node_loss "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/training/loss/DVH_loss.py"
click node_dataloader "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/training/dataloading/base_data_loader.py"
click node_predictor "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/inference/predict_from_raw_data.py"
click node_iterator "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/inference/data_iterators.py"
click node_sliding "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/inference/sliding_window_prediction.py"
click node_export "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/inference/export_prediction.py"
click node_evaluation "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/evaluation/evaluate_predictions.py"
click node_dicom_eval "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/evaluation/dicom_based_evaluation.py"
click node_dicom_out "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/dicom_bridge/nifti2dicom_dose.py"
click node_gamma "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/evaluation/gamma_analysis.py"
click node_ensemble "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/ensembling/ensemble.py"
click node_imageio "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/imageio/reader_writer_registry.py"
click node_reader "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/imageio/simpleitk_reader_writer.py"
click node_plans "https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/utilities/plans_handling/plans_handler.py"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_format,node_fingerprint,node_preprocess,node_resample toneBlue
class node_planning,node_plan_api,node_trainer,node_train_run,node_loss,node_dataloader toneAmber
class node_predictor,node_iterator,node_sliding,node_export,node_evaluation,node_dicom_eval,node_gamma,node_ensemble toneMint
class node_dicom_in,node_dicom_out,node_imageio,node_reader,node_plans toneRose
class node_user toneIndigo
```

## Components

### Dataset preparation

- [Format and label](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/formating_dataset/format_dataset_api.py): `nndosenet/formating_dataset/format_dataset_api.py`
- [Dataset fingerprint](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/experiment_planning/dataset_fingerprint/fingerprint_extractor.py): `nndosenet/experiment_planning/dataset_fingerprint/fingerprint_extractor.py`
- [Preprocessing](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/preprocessing/preprocessors/default_preprocessor.py): `nndosenet/preprocessing/preprocessors/default_preprocessor.py`
- [Resampling](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/preprocessing/resampling/default_resampling.py): `nndosenet/preprocessing/resampling/default_resampling.py`

### Planning and training

- [Experiment planning](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/experiment_planning/experiment_planners/default_experiment_planner.py): `nndosenet/experiment_planning/experiment_planners/default_experiment_planner.py`
- [Plan entrypoints](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/experiment_planning/plan_and_preprocess_entrypoints.py): `nndosenet/experiment_planning/plan_and_preprocess_entrypoints.py`
- [Dose trainer](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/training/nnDoseTrainer/nnDoseTrainer.py): `nndosenet/training/nnDoseTrainer/nnDoseTrainer.py`
- [Training runner](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/run/run_training.py): `nndosenet/run/run_training.py`
- [DVH loss](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/training/loss/DVH_loss.py): `nndosenet/training/loss/DVH_loss.py`
- [Training data loading](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/training/dataloading/base_data_loader.py): `nndosenet/training/dataloading/base_data_loader.py`

### Inference and evaluation

- [Prediction runner](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/inference/predict_from_raw_data.py): `nndosenet/inference/predict_from_raw_data.py`
- [Prediction data iterator](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/inference/data_iterators.py): `nndosenet/inference/data_iterators.py`
- [Sliding-window prediction](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/inference/sliding_window_prediction.py): `nndosenet/inference/sliding_window_prediction.py`
- [Prediction export](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/inference/export_prediction.py): `nndosenet/inference/export_prediction.py`
- [Dose evaluation](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/evaluation/evaluate_predictions.py): `nndosenet/evaluation/evaluate_predictions.py`
- [DICOM evaluation](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/evaluation/dicom_based_evaluation.py): `nndosenet/evaluation/dicom_based_evaluation.py`
- [Gamma analysis](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/evaluation/gamma_analysis.py): `nndosenet/evaluation/gamma_analysis.py`
- [Model ensembling](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/ensembling/ensemble.py): `nndosenet/ensembling/ensemble.py`

### Medical imaging support

- [DICOM conversion](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/dicom_bridge/dicom2nifti.py): `nndosenet/dicom_bridge/dicom2nifti.py`
- [NIfTI to DICOM](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/dicom_bridge/nifti2dicom_dose.py): `nndosenet/dicom_bridge/nifti2dicom_dose.py`
- [Image I/O registry](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/imageio/reader_writer_registry.py): `nndosenet/imageio/reader_writer_registry.py`
- [SimpleITK reader](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/imageio/simpleitk_reader_writer.py): `nndosenet/imageio/simpleitk_reader_writer.py`
- [Plans and labels](https://github.com/rita12321/nndosenet/blob/nnDoseNet/nndosenet/utilities/plans_handling/plans_handler.py): `nndosenet/utilities/plans_handling/plans_handler.py`

### Other components

- **Researcher**

## Connections

- Researcher → Format and label: provides cases
- Format and label → DICOM conversion: converts DICOM (evidence: `nndosenet/formating_dataset/format_dataset_api.py`)
- Researcher → Plan entrypoints: requests planning
- Plan entrypoints → Experiment planning: runs planner (evidence: `nndosenet/experiment_planning/plan_and_preprocess_entrypoints.py`)
- Experiment planning → Dataset fingerprint: analyzes dataset (evidence: `nndosenet/experiment_planning/experiment_planners/default_experiment_planner.py`)
- Plan entrypoints → Preprocessing: preprocesses cases
- Preprocessing → Resampling: resamples images (evidence: `nndosenet/preprocessing/preprocessors/default_preprocessor.py`)
- Researcher → Training runner: starts training
- Training runner → Dose trainer: instantiates trainer (evidence: `nndosenet/run/run_training.py`)
- Dose trainer → Training data loading: loads batches (evidence: `nndosenet/training/nnDoseTrainer/nnDoseTrainer.py`)
- Dose trainer → DVH loss: computes DVH loss (evidence: `nndosenet/training/nnDoseTrainer/nnDoseTrainer.py`)
- Dose trainer → Dose evaluation: evaluates validation (evidence: `nndosenet/training/nnDoseTrainer/nnDoseTrainer.py`)
- Researcher → Prediction runner: requests prediction
- Prediction runner → Prediction data iterator: iterates inputs (evidence: `nndosenet/inference/predict_from_raw_data.py`)
- Prediction runner → Sliding-window prediction: predicts patches (evidence: `nndosenet/inference/predict_from_raw_data.py`)
- Prediction runner → Prediction export: exports predictions (evidence: `nndosenet/inference/predict_from_raw_data.py`)
- Prediction runner → Dose evaluation: computes metrics (evidence: `nndosenet/inference/predict_from_raw_data.py`)
- Dose evaluation → Image I/O registry: selects reader (evidence: `nndosenet/evaluation/evaluate_predictions.py`)
- Image I/O registry → SimpleITK reader: registers reader (evidence: `nndosenet/imageio/reader_writer_registry.py`)
- Prediction runner → Plans and labels: loads plan metadata (evidence: `nndosenet/inference/predict_from_raw_data.py`)
- Researcher → DICOM evaluation: requests RT analysis
- DICOM evaluation → NIfTI to DICOM: converts dose (evidence: `nndosenet/evaluation/dicom_based_evaluation.py`)
- DICOM evaluation → Gamma analysis: runs gamma analysis (evidence: `nndosenet/evaluation/dicom_based_evaluation.py`)
- Researcher → Model ensembling: combines models
- Model ensembling → Dose evaluation: evaluates ensemble (evidence: `nndosenet/ensembling/ensemble.py`)
