# mostly-ai/paper-tabular-argn architecture

> Architecture diagram of the GitHub repository mostly-ai/paper-tabular-argn, 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/mostly-ai/paper-tabular-argn
- Repository: https://github.com/mostly-ai/paper-tabular-argn

## Overview

This repository is a research artifact for the TabularARGN paper: it provides training datasets, scripts for generating synthetic tabular data with TabularARGN and comparison methods, and generated outputs. The principal workflow is a researcher running a method script on training data and saving synthetic datasets. The excerpts show flat and relational generation separately. Source connections below are limited to sampled scripts; the README identifies the paper and points to the reference implementation in another repository, which is outside this tree.

Main entry points: [`data_train`](https://github.com/mostly-ai/paper-tabular-argn/tree/main/data_train) (Training data), [`data_synthetic`](https://github.com/mostly-ai/paper-tabular-argn/tree/main/data_synthetic) (Generated datasets), [`scripts_flat/run_tabARGN.py`](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_flat/run_tabARGN.py) (TabularARGN flat), [`scripts_sequential/script_tabARGN_california.py`](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_tabARGN_california.py) (TabularARGN relational)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_data["Input datasets"]
  node_train_data[("Training data")]
end

subgraph group_flat["Flat generation"]
  node_argn_flat["TabularARGN flat<br/>[run_tabARGN.py]"]
  node_flat_baselines["Flat baselines"]
end

subgraph group_relational["Relational generation"]
  node_argn_rel["TabularARGN relational"]
  node_argn_baseball["TabularARGN baseball"]
  node_rctgan["RCTGAN"]
  node_rctgan_baseball["RCTGAN baseball"]
  node_rtf["REaLTabFormer"]
  node_rtf_baseball["REaLTabFormer baseball"]
  node_sdv_hma["SDV HMA"]
  node_sdv_hma_baseball["SDV HMA baseball"]
  node_sdv_par["SDV PAR"]
  node_sdv_par_baseball["SDV PAR baseball"]
end

subgraph group_results["Synthetic outputs"]
  node_synthetic[("Generated datasets")]
end

node_researcher(("Researcher"))

node_researcher -->|"runs"| node_argn_flat
node_researcher -->|"runs"| node_flat_baselines
node_researcher -->|"runs"| node_argn_rel
node_researcher -->|"runs"| node_argn_baseball
node_researcher -->|"runs"| node_rctgan
node_researcher -->|"runs"| node_rctgan_baseball
node_researcher -->|"runs"| node_rtf
node_researcher -->|"runs"| node_rtf_baseball
node_researcher -->|"runs"| node_sdv_hma
node_researcher -->|"runs"| node_sdv_hma_baseball
node_researcher -->|"runs"| node_sdv_par
node_researcher -->|"runs"| node_sdv_par_baseball
node_argn_flat -->|"reads"| node_train_data
node_flat_baselines -->|"preprocesses"| node_train_data
node_argn_rel -->|"reads"| node_train_data
node_argn_baseball -->|"reads"| node_train_data
node_rctgan -->|"reads"| node_train_data
node_rctgan_baseball -->|"reads"| node_train_data
node_rtf -->|"reads"| node_train_data
node_rtf_baseball -->|"reads"| node_train_data
node_sdv_hma -->|"reads"| node_train_data
node_sdv_hma_baseball -->|"reads"| node_train_data
node_sdv_par -->|"reads"| node_train_data
node_sdv_par_baseball -->|"reads"| node_train_data
node_argn_flat -->|"writes"| node_synthetic
node_argn_rel -->|"writes"| node_synthetic
node_argn_baseball -->|"writes"| node_synthetic
node_rctgan -->|"writes"| node_synthetic
node_rctgan_baseball -->|"writes"| node_synthetic
node_rtf -->|"writes"| node_synthetic
node_rtf_baseball -->|"writes"| node_synthetic
node_sdv_hma -->|"writes"| node_synthetic
node_sdv_hma_baseball -->|"writes"| node_synthetic
node_sdv_par -->|"writes"| node_synthetic
node_sdv_par_baseball -->|"writes"| node_synthetic

click node_train_data "https://github.com/mostly-ai/paper-tabular-argn/tree/main/data_train"
click node_argn_flat "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_flat/run_tabARGN.py"
click node_flat_baselines "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_flat/run_baselines_flat.py"
click node_argn_rel "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_tabARGN_california.py"
click node_argn_baseball "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_tabARGN_baseball.py"
click node_rctgan "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_rctgan_california.py"
click node_rctgan_baseball "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_rctgan_baseball.py"
click node_rtf "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_rtf_california.py"
click node_rtf_baseball "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_rtf_baseball.py"
click node_sdv_hma "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_sdv_hma_california.py"
click node_sdv_hma_baseball "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_sdv_hma_baseball.py"
click node_sdv_par "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_sdv_par_california.py"
click node_sdv_par_baseball "https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_sdv_par_baseball.py"
click node_synthetic "https://github.com/mostly-ai/paper-tabular-argn/tree/main/data_synthetic"

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_train_data toneBlue
class node_argn_flat,node_flat_baselines toneAmber
class node_argn_rel,node_argn_baseball,node_rctgan,node_rctgan_baseball,node_rtf,node_rtf_baseball,node_sdv_hma,node_sdv_hma_baseball,node_sdv_par,node_sdv_par_baseball toneMint
class node_synthetic toneRose
class node_researcher toneIndigo
```

## Components

### Input datasets

- [Training data](https://github.com/mostly-ai/paper-tabular-argn/tree/main/data_train): `data_train`

### Flat generation

- [TabularARGN flat](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_flat/run_tabARGN.py): `scripts_flat/run_tabARGN.py`
- [Flat baselines](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_flat/run_baselines_flat.py): `scripts_flat/run_baselines_flat.py`

### Relational generation

- [TabularARGN relational](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_tabARGN_california.py): `scripts_sequential/script_tabARGN_california.py`
- [TabularARGN baseball](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_tabARGN_baseball.py): `scripts_sequential/script_tabARGN_baseball.py`
- [RCTGAN](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_rctgan_california.py): `scripts_sequential/script_rctgan_california.py`
- [RCTGAN baseball](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_rctgan_baseball.py): `scripts_sequential/script_rctgan_baseball.py`
- [REaLTabFormer](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_rtf_california.py): `scripts_sequential/script_rtf_california.py`
- [REaLTabFormer baseball](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_rtf_baseball.py): `scripts_sequential/script_rtf_baseball.py`
- [SDV HMA](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_sdv_hma_california.py): `scripts_sequential/script_sdv_hma_california.py`
- [SDV HMA baseball](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_sdv_hma_baseball.py): `scripts_sequential/script_sdv_hma_baseball.py`
- [SDV PAR](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_sdv_par_california.py): `scripts_sequential/script_sdv_par_california.py`
- [SDV PAR baseball](https://github.com/mostly-ai/paper-tabular-argn/blob/main/scripts_sequential/script_sdv_par_baseball.py): `scripts_sequential/script_sdv_par_baseball.py`

### Synthetic outputs

- [Generated datasets](https://github.com/mostly-ai/paper-tabular-argn/tree/main/data_synthetic): `data_synthetic`

### Other components

- **Researcher**

## Connections

- Researcher → TabularARGN flat: runs (evidence: `scripts_flat/run_tabARGN.py`)
- Researcher → Flat baselines: runs (evidence: `scripts_flat/run_baselines_flat.py`)
- Researcher → TabularARGN relational: runs (evidence: `scripts_sequential/script_tabARGN_california.py`)
- Researcher → TabularARGN baseball: runs (evidence: `scripts_sequential/script_tabARGN_baseball.py`)
- Researcher → RCTGAN: runs (evidence: `scripts_sequential/script_rctgan_california.py`)
- Researcher → RCTGAN baseball: runs (evidence: `scripts_sequential/script_rctgan_baseball.py`)
- Researcher → REaLTabFormer: runs (evidence: `scripts_sequential/script_rtf_california.py`)
- Researcher → REaLTabFormer baseball: runs (evidence: `scripts_sequential/script_rtf_baseball.py`)
- Researcher → SDV HMA: runs (evidence: `scripts_sequential/script_sdv_hma_california.py`)
- Researcher → SDV HMA baseball: runs (evidence: `scripts_sequential/script_sdv_hma_baseball.py`)
- Researcher → SDV PAR: runs (evidence: `scripts_sequential/script_sdv_par_california.py`)
- Researcher → SDV PAR baseball: runs (evidence: `scripts_sequential/script_sdv_par_baseball.py`)
- TabularARGN flat → Training data: reads (evidence: `scripts_flat/run_tabARGN.py`)
- Flat baselines → Training data: preprocesses (evidence: `scripts_flat/run_baselines_flat.py`)
- TabularARGN relational → Training data: reads (evidence: `scripts_sequential/script_tabARGN_california.py`)
- TabularARGN baseball → Training data: reads (evidence: `scripts_sequential/script_tabARGN_baseball.py`)
- RCTGAN → Training data: reads (evidence: `scripts_sequential/script_rctgan_california.py`)
- RCTGAN baseball → Training data: reads (evidence: `scripts_sequential/script_rctgan_baseball.py`)
- REaLTabFormer → Training data: reads (evidence: `scripts_sequential/script_rtf_california.py`)
- REaLTabFormer baseball → Training data: reads (evidence: `scripts_sequential/script_rtf_baseball.py`)
- SDV HMA → Training data: reads (evidence: `scripts_sequential/script_sdv_hma_california.py`)
- SDV HMA baseball → Training data: reads (evidence: `scripts_sequential/script_sdv_hma_baseball.py`)
- SDV PAR → Training data: reads (evidence: `scripts_sequential/script_sdv_par_california.py`)
- SDV PAR baseball → Training data: reads (evidence: `scripts_sequential/script_sdv_par_baseball.py`)
- TabularARGN flat → Generated datasets: writes (evidence: `scripts_flat/run_tabARGN.py`)
- TabularARGN relational → Generated datasets: writes (evidence: `scripts_sequential/script_tabARGN_california.py`)
- TabularARGN baseball → Generated datasets: writes (evidence: `scripts_sequential/script_tabARGN_baseball.py`)
- RCTGAN → Generated datasets: writes (evidence: `scripts_sequential/script_rctgan_california.py`)
- RCTGAN baseball → Generated datasets: writes (evidence: `scripts_sequential/script_rctgan_baseball.py`)
- REaLTabFormer → Generated datasets: writes (evidence: `scripts_sequential/script_rtf_california.py`)
- REaLTabFormer baseball → Generated datasets: writes (evidence: `scripts_sequential/script_rtf_baseball.py`)
- SDV HMA → Generated datasets: writes (evidence: `scripts_sequential/script_sdv_hma_california.py`)
- SDV HMA baseball → Generated datasets: writes (evidence: `scripts_sequential/script_sdv_hma_baseball.py`)
- SDV PAR → Generated datasets: writes (evidence: `scripts_sequential/script_sdv_par_california.py`)
- SDV PAR baseball → Generated datasets: writes (evidence: `scripts_sequential/script_sdv_par_baseball.py`)
