# palevas/syntax-score-prediction architecture

> Architecture diagram of the GitHub repository palevas/syntax-score-prediction, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-01). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/palevas/syntax-score-prediction
- Repository: https://github.com/palevas/syntax-score-prediction

## Overview

This repository implements an end-to-end video-based SYNTAX score prediction benchmark. Its workflow trains an artery-specific video backbone, uses that backbone to initialize a study-level sequence model, then applies sequence models to left and right coronary artery videos and combines their estimates into a study-level prediction. The dataset classes read study metadata and DICOM video. README and sampled source establish the training and inference paths; other benchmark capabilities, such as model variants, are retained, but unsampled internals are not expanded.

Main entry points: [`seq_apply.py`](https://github.com/palevas/syntax-score-prediction/blob/main/seq_apply.py) (Inference runner), [`seq_train.py`](https://github.com/palevas/syntax-score-prediction/blob/main/seq_train.py) (Sequence training), [`backbone_train.py`](https://github.com/palevas/syntax-score-prediction/blob/main/backbone_train.py) (Backbone training), [`seq_dataset.py`](https://github.com/palevas/syntax-score-prediction/blob/main/seq_dataset.py) (Study video dataset)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_backbone["Clip backbone"]
  node_backbone_data["Clip dataset"]
  node_backbone_training["Backbone training<br/>[backbone_train.py]"]
  node_backbone_model["Video classifier<br/>[backbone_model.py]"]
  node_backbone_weights["Backbone weights"]
end

subgraph group_sequence["Study sequence"]
  node_sequence_data["Study video dataset<br/>[seq_dataset.py]"]
  node_sequence_training["Sequence training<br/>[seq_train.py]"]
  node_sequence_model["Sequence predictor<br/>[seq_model.py]"]
  node_sequence_weights["Sequence weights"]
end

subgraph group_inference["Prediction"]
  node_inference_runner["Inference runner<br/>[seq_apply.py]"]
  node_study_predictions["Study predictions"]
  node_metrics["Evaluation metrics<br/>[seq_apply.py]"]
end

node_user(("Researcher"))
node_dataset["CardioSYNTAX data"]

node_user -->|"trains backbone"| node_backbone_training
node_dataset -->|"provides metadata and video"| node_backbone_data
node_backbone_training -->|"loads batches"| node_backbone_data
node_backbone_training -->|"trains model"| node_backbone_model
node_backbone_training -->|"saves weights"| node_backbone_weights
node_user -->|"trains sequence model"| node_sequence_training
node_dataset -->|"provides study videos"| node_sequence_data
node_sequence_training -->|"loads batches"| node_sequence_data
node_sequence_training -->|"trains model"| node_sequence_model
node_sequence_training -->|"loads initialization"| node_backbone_weights
node_sequence_training -->|"saves checkpoints"| node_sequence_weights
node_user -->|"runs inference"| node_inference_runner
node_dataset -->|"provides study metadata"| node_inference_runner
node_inference_runner -->|"loads videos"| node_sequence_data
node_inference_runner -->|"runs prediction"| node_sequence_model
node_sequence_weights -->|"provides model weights"| node_inference_runner
node_inference_runner -->|"writes study scores"| node_study_predictions
node_inference_runner -->|"computes metrics"| node_metrics

click node_backbone_data "https://github.com/palevas/syntax-score-prediction/blob/main/backbone_dataset.py"
click node_backbone_training "https://github.com/palevas/syntax-score-prediction/blob/main/backbone_train.py"
click node_backbone_model "https://github.com/palevas/syntax-score-prediction/blob/main/backbone_model.py"
click node_sequence_data "https://github.com/palevas/syntax-score-prediction/blob/main/seq_dataset.py"
click node_sequence_training "https://github.com/palevas/syntax-score-prediction/blob/main/seq_train.py"
click node_sequence_model "https://github.com/palevas/syntax-score-prediction/blob/main/seq_model.py"
click node_inference_runner "https://github.com/palevas/syntax-score-prediction/blob/main/seq_apply.py"
click node_metrics "https://github.com/palevas/syntax-score-prediction/blob/main/seq_apply.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_backbone_data,node_backbone_training,node_backbone_model,node_backbone_weights toneBlue
class node_sequence_data,node_sequence_training,node_sequence_model,node_sequence_weights toneAmber
class node_inference_runner,node_study_predictions,node_metrics toneMint
class node_user,node_dataset toneIndigo
```

## Components

### Clip backbone

- [Clip dataset](https://github.com/palevas/syntax-score-prediction/blob/main/backbone_dataset.py): `backbone_dataset.py`
- [Backbone training](https://github.com/palevas/syntax-score-prediction/blob/main/backbone_train.py): `backbone_train.py`
- [Video classifier](https://github.com/palevas/syntax-score-prediction/blob/main/backbone_model.py): `backbone_model.py`
- **Backbone weights**

### Study sequence

- [Study video dataset](https://github.com/palevas/syntax-score-prediction/blob/main/seq_dataset.py): `seq_dataset.py`
- [Sequence training](https://github.com/palevas/syntax-score-prediction/blob/main/seq_train.py): `seq_train.py`
- [Sequence predictor](https://github.com/palevas/syntax-score-prediction/blob/main/seq_model.py): `seq_model.py`
- **Sequence weights**

### Prediction

- [Inference runner](https://github.com/palevas/syntax-score-prediction/blob/main/seq_apply.py): `seq_apply.py`
- **Study predictions**
- [Evaluation metrics](https://github.com/palevas/syntax-score-prediction/blob/main/seq_apply.py): `seq_apply.py`

### Other components

- **Researcher**
- **CardioSYNTAX data**

## Connections

- Researcher → Backbone training: trains backbone (evidence: `README.md`)
- CardioSYNTAX data → Clip dataset: provides metadata and video (evidence: `backbone_dataset.py`)
- Backbone training → Clip dataset: loads batches (evidence: `backbone_train.py`)
- Backbone training → Video classifier: trains model (evidence: `backbone_train.py`)
- Backbone training → Backbone weights: saves weights (evidence: `backbone_train.py`)
- Researcher → Sequence training: trains sequence model (evidence: `README.md`)
- CardioSYNTAX data → Study video dataset: provides study videos (evidence: `seq_dataset.py`)
- Sequence training → Study video dataset: loads batches (evidence: `seq_train.py`)
- Sequence training → Sequence predictor: trains model (evidence: `seq_train.py`)
- Sequence training → Backbone weights: loads initialization (evidence: `seq_train.py`)
- Sequence training → Sequence weights: saves checkpoints (evidence: `seq_train.py`)
- Researcher → Inference runner: runs inference (evidence: `README.md`)
- CardioSYNTAX data → Inference runner: provides study metadata (evidence: `seq_apply.py`)
- Inference runner → Study video dataset: loads videos (evidence: `seq_apply.py`)
- Inference runner → Sequence predictor: runs prediction (evidence: `seq_apply.py`)
- Sequence weights → Inference runner: provides model weights (evidence: `seq_apply.py`)
- Inference runner → Study predictions: writes study scores (evidence: `seq_apply.py`)
- Inference runner → Evaluation metrics: computes metrics (evidence: `seq_apply.py`)
