Metadata-Version: 2.5
Name: dp-cgans
Version: 0.3.0
Summary: A library to generate synthetic tabular or RDF data using Conditional Generative Adversary Networks (GANs) combined with Differential Privacy techniques.
Project-URL: Homepage, https://github.com/sunchang0124/dp_cgans
Project-URL: Documentation, https://github.com/sunchang0124/dp_cgans
Project-URL: History, https://github.com/sunchang0124/dp_cgans/releases
Project-URL: Tracker, https://github.com/sunchang0124/dp_cgans/issues
Project-URL: Source, https://github.com/sunchang0124/dp_cgans
Author-email: Sun Chang <sunchang0124@gmail.com>
Maintainer-email: Sun Chang <sunchang0124@gmail.com>, Vincent Emonet <vincent.emonet@gmail.com>
License: MIT License
        
        Copyright (c) 2025-present Sun Chang <chang.sun@maastrichtuniversity.nl>
        
        Permission is hereby granted, free of charge, to any person obtaining a copy
        of this software and associated documentation files (the "Software"), to deal
        in the Software without restriction, including without limitation the rights
        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
        copies of the Software, and to permit persons to whom the Software is
        furnished to do so, subject to the following conditions:
        
        The above copyright notice and this permission notice shall be included in all
        copies or substantial portions of the Software.
        
        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
        SOFTWARE.
License-File: LICENSE.txt
Keywords: CGAN,DP,Differential Privacy,GAN,synthetic data
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Requires-Python: >=3.10
Requires-Dist: copulas
Requires-Dist: gensim
Requires-Dist: graphviz
Requires-Dist: matplotlib
Requires-Dist: numpy
Requires-Dist: pandas
Requires-Dist: pyreadstat
Requires-Dist: scikit-learn
Requires-Dist: scipy
Requires-Dist: torch
Requires-Dist: tqdm
Requires-Dist: typer
Provides-Extra: test
Requires-Dist: black; extra == 'test'
Requires-Dist: mypy>=1.4.1; extra == 'test'
Requires-Dist: pip-tools; extra == 'test'
Requires-Dist: pytest-cov>=3.0.0; extra == 'test'
Requires-Dist: pytest>=7.4.0; extra == 'test'
Requires-Dist: ruff; extra == 'test'
Requires-Dist: types-pyyaml; extra == 'test'
Requires-Dist: types-requests; extra == 'test'
Description-Content-Type: text/markdown

# 👯 DP-CGANS (Differentially Private - Conditional Generative Adversarial NetworkS)

[![PyPi Shield](https://img.shields.io/pypi/v/dp-cgans)](https://pypi.org/project/dp-cgans/) [![Py versions](https://img.shields.io/pypi/pyversions/dp-cgans)](https://pypi.org/project/dp-cgans/) [![Test package](https://github.com/sunchang0124/dp_cgans/actions/workflows/test.yml/badge.svg)](https://github.com/sunchang0124/dp_cgans/actions/workflows/test.yml) [![Publish package](https://github.com/sunchang0124/dp_cgans/actions/workflows/publish.yml/badge.svg)](https://github.com/sunchang0124/dp_cgans/actions/workflows/publish.yml)



<!-- [![Development Status](https://img.shields.io/badge/Development%20Status-2%20--%20Pre--Alpha-yellow)](https://pypi.org/search/?c=Development+Status+%3A%3A+2+-+Pre-Alpha) -->
<!-- [![PyPi Shield](https://img.shields.io/badge/pypi-v0.0.2-blue)](https://pypi.org/project/dp-cgans/) -->
<!-- [![Tests](https://github.com/sdv-dev/SDV/workflows/Run%20Tests/badge.svg)](https://github.com/sdv-dev/SDV/actions?query=workflow%3A%22Run+Tests%22+branch%3Amaster) -->

**Abstract**: This repository presents a Conditional Generative Adversary Networks (GANs) on tabular data (and RDF data) combining with Differential Privacy techniques. Our pre-print publication: [Generating synthetic personal health data using conditional generative adversarial networks combining with differential privacy](https://doi.org/10.1016/j.jbi.2023.104404).

**Author**: Chang Sun, Institute of Data Science, Maastricht University
**Start date**: Nov-2021
**Status**: Under development

**Note**: "Standing on the shoulders of giants". This repository is inspired by the excellent work of [CTGAN](https://github.com/sdv-dev/CTGAN) from [Synthetic Data Vault (SDV)](https://github.com/sdv-dev/SDV), [Tensorflow Privacy](https://github.com/tensorflow/privacy), and [RdfPdans](https://github.com/cadmiumkitty/rdfpandas). Highly appreciate they shared the ideas and implementations, made code publicly available, well-written documentation. More related work can be found in the References below.  

This package is extended from SDV (https://github.com/sdv-dev/SDV), CTGAN (https://github.com/sdv-dev/CTGAN), and Differential Privacy in GANs (https://github.com/civisanalytics/dpwgan). The author modified the conditional matrix and cost functions to emphasize on the relations between variables. The main changes are in ctgan/synthesizers/ctgan.py ../data_sampler.py ../data_transformer.py


## 🎯 What to use DP-CGANS for

DP-CGANS learns the joint distribution of a tabular dataset and generates new rows that follow it, optionally under a differential-privacy guarantee. The table lists the purposes the model is built for and how the package supports each one.

| Purpose | What it means | How DP-CGANS supports it |
|---|---|---|
| **Open release** | The synthetic output is shared or published in place of the real data. | Train with `private=True`: the discriminator is trained with DP-SGD and the spent privacy budget (ε, δ) is reported with Rényi DP accounting, so the generated data can be released as a privacy-preserving substitute. |
| **Pipeline testing** | Exercising software, formats or pipelines; no realism claim. | Any CSV goes in, a CSV with the same columns and types comes out. `dp-cgans gen` produces development data in one command. |
| **ML augmentation** | Adding synthetic rows to train a predictive model. | Sample as many rows as you need with `model.sample(n)`. The paper evaluates train-on-synthetic, test-on-real performance. |
| **Statistical replication** | The synthetic data reproduces the real distributions and relationships closely enough to rerun an analysis on it. | The conditional vector and loss functions were modified to emphasise the relations between variables. The paper evaluates marginal distributions, pairwise correlations and statistical similarity against the real data. |
| **Causal inference** | The synthetic data preserves a known or specified causal structure. | DP-CGANS fits a joint distribution; it does not encode a causal graph. Use it to exercise a causal-analysis pipeline, not to validate causal claims. |
| **Education** | Synthetic data used as teaching material, for demonstrations or in training courses. | Generate a realistic but non-identifiable copy of a sensitive dataset so students can practise on it, for example medical students learning statistical analysis on synthetic patient records. Trained with `private=True`, the teaching data carries a formal privacy guarantee and can be handed out freely. |
| **Benchmarking** | Used as a baseline or benchmark dataset for comparing generators or metrics. | A published method with a pip-installable implementation, a fixed example dataset and a CLI, so it can serve as a reproducible baseline in generator comparisons. |
| **Imbalance correction** | Oversampling a rare class. | `model.sample(n, condition_column=..., condition_value=...)` draws rows for a chosen category, so a minority class can be oversampled directly. |

## 📥️ Installation

You will need Python >=3.8+

```shell
pip3 install -U dp-cgans
```

## 🪄 Usage

### 🐍 Use with Python

Use the library directly in Python scripts when you want full control over the model hyper-parameters, the differential-privacy setting (`private=True`), and conditional sampling.

New to DP-CGANS? Try it first in the [tutorial notebook](https://colab.research.google.com/github/sunchang0124/syndata-workshop/blob/main/dpcgans_hands_on.ipynb) on Colab, which runs the example below step by step in your browser.

If your input is tabular data (e.g., csv):

 ```python
from dp_cgans import DP_CGAN
import pandas as pd
import time

input_file = "../resources/example_tabular_data_UCIAdult.csv"
tabular_data = pd.read_csv(input_file)

### Add your pre-processing if needed
for col in tabular_data.columns:
   tabular_data[col] = pd.to_numeric(tabular_data[col], errors='ignore', downcast='integer')
for col in tabular_data.columns:
   if tabular_data[col].nunique() < 10: 
       tabular_data[col] = tabular_data[col].astype('object')

# Configure model hyper-parameters
model = DP_CGAN(
   epochs=100, # number of training epochs
   batch_size=100, # the size of each batch
   log_frequency=True,
   verbose=True,
   generator_dim=(128, 128, 128),
   discriminator_dim=(128, 128, 128),
   generator_lr=2e-4, 
   discriminator_lr=2e-4,
   discriminator_steps=10, 
   private=False,
   saved_transformer=None #"fitted_transformer.pkl"
)

start_time = time.time()
print("Start training the model: ")
model.fit(tabular_data)
end_time = time.time()

elapsed_time = end_time - start_time
print("Training model time ", elapsed_time)

print("Saving the trained generator...")
model.save("generator.pkl")

# print("load the trained file.")
# loaded_model=DP_CGAN.load("PATH_TO_MODEL")

# Generate 100 synthetic rows.
# Numeric columns come back with the same dtype and the same number of
# decimal places as the real data (e.g. 2 decimals in, 2 decimals out).
syn_data = model.sample(100)

syn_data.to_csv("syn_data_file.csv",index=None)
 ```

<!-- 
2. If your input data is in RDF format:

  ```python
from dp_cgans import DP_CGAN
from dp_cgans import RDF_to_Tabular

# Step 1. Load RDF to a plain table (dataframe)
plain_tabular=RDF_to_Tabular(file_path="../resources/example_rdf_data.ttl")

# Step 2. Convert plain table to a structured table 
# After step 1, RDF data will be converted a plain tabular dataset (all the nodes/entities will be presented as rows. Step 2 will structure the table by recognizing and sorting the types of the entities, replacing the URI with actual value which is attached to that URI. Users can decide how many levels they want to unfold their RDF models to tabular datasets.)
tabular_data,rel_pred_obj=plain_tabular.fit_convert(user_define_data_instance="http://ncicb.nci.nih.gov/xml/owl/EVS/Thesaurus.owl#C16960", 
                                                    user_define_is_a=["rdf:type{URIRef}"], 
                                                    user_define_has_value=["http://www.cancerdata.org/roo/P100042"], 
                                                    set_level="full", 
                                                    as_column='object', 
                                                    remove_columns_unique_values=True)

# Step 3. Define your GANS model
model = DP_CGAN(
    epochs=100, # number of training epochs
    batch_size=1000, # the size of each batch
    log_frequency=True,
    verbose=True,
    generator_dim=(128, 128, 128),
    discriminator_dim=(128, 128, 128),
    generator_lr=2e-4, 
    discriminator_lr=2e-4,
    discriminator_steps=1, 
    private=False,
)

print("Start training model")
model.fit(tabular_data)

# Sample the generated synthetic data
model.sample(100)
  ```
-->

### ⌨️ Use as a command-line interface

If you only need a quick synthetic copy of a CSV file, you can generate it from your terminal after installing `dp-cgans` with pip (default hyper-parameters, no differential privacy).

To quickly run our example, you can download the [example data](https://raw.githubusercontent.com/sunchang0124/dp_cgans/main/resources/example_tabular_data_UCIAdult.csv):

```bash
wget https://raw.githubusercontent.com/sunchang0124/dp_cgans/main/resources/example_tabular_data_UCIAdult.csv
```

Then run `dp-cgans`:

```bash
dp-cgans gen example_tabular_data_UCIAdult.csv --epochs 100 --output out.csv --gen-size 100
```

Get a full rundown of the available options for generating synthetic data with:

```bash
dp-cgans gen --help
```

### 📓 Tutorial notebook

A hands-on tutorial that walks through loading a dataset, training DP-CGANS, sampling synthetic rows, and inspecting the result runs in the browser, no installation needed:

[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/sunchang0124/syndata-workshop/blob/main/dpcgans_hands_on.ipynb) **DP-CGANS hands-on notebook** (source: [sunchang0124/syndata-workshop](https://github.com/sunchang0124/syndata-workshop/blob/main/dpcgans_hands_on.ipynb))


## 🧑‍💻 Development setup


For development, we recommend to install and use [Hatch](https://hatch.pypa.io/latest/), as it will automatically install and sync the dependencies when running development scripts. But you can also directly create a virtual environment and install the library with `pip install -e .`

### Install

Clone the repository:

```bash
git clone https://github.com/sunchang0124/dp_cgans
cd dp_cgans
```

> When working in development the `hatch` tool will automatically install and sync the dependencies when running a script. But you can also directly 

### Run

Run the library with the CLI:

```bash
hatch -v run dp-cgans gen --help
```

You can also enter a new shell with the virtual environments automatically activated:

```bash
hatch shell
dp-cgans gen --help
```

### Tests

Run the tests locally:

```bash
hatch run pytest -s
```

### Format

Run formatting and linting (black and ruff):

```bash
hatch run fmt
```

### Reset the virtual environments

In case the virtual environments is not updating as expected you can easily reset it with:

```bash
hatch env prune
```

## 📦️ New release process

The deployment of new releases is done automatically by a GitHub Action workflow when a new release is created on GitHub. To release a new version:

1. Make sure the `PYPI_API_TOKEN` secret has been defined in the GitHub repository (in Settings > Secrets > Actions). You can get an API token from PyPI [here](https://pypi.org/manage/account/).

2. Increment the `version` number in `src/dp_cgans/__init__.py` file:

   ```bash
   hatch version fix    # Bump from 0.0.1 to 0.0.2
   hatch version minor  # Bump from 0.0.1 to 0.1.0
   hatch version 0.1.1  # Bump to the specified version
   ```

3. Create a new release on GitHub, which will automatically trigger the publish workflow, and publish the new release to PyPI.

You can also manually build and publish from you laptop:

```bash
hatch build
hatch publish
```

## 📚️ References / Further reading 

There are many excellent work on generating synthetic data using GANS and other methods. We list the studies that made great conbributions for the field and inspiring for our work.

##### GANS

   1. Synthetic Data Vault (SDV) [[Paper](https://dai.lids.mit.edu/wp-content/uploads/2018/03/SDV.pdf)] [[Github](https://github.com/sdv-dev/SDV)]
   2. Modeling Tabular Data using Conditional GAN (a part of SDV) [[Paper](https://arxiv.org/abs/1907.00503)] [[Github](https://github.com/sdv-dev/CTGAN)]
   3. Wasserstein GAN [[Paper](https://arxiv.org/pdf/1701.07875.pdf)]
   4. Improved Training of Wasserstein GANs [[Paper](https://papers.nips.cc/paper/2017/file/892c3b1c6dccd52936e27cbd0ff683d6-Paper.pdf)]
   5. Synthesising Tabular Data using Wasserstein Conditional GANs with Gradient Penalty (WCGAN-GP) [[Paper](http://ceur-ws.org/Vol-2771/AICS2020_paper_57.pdf)]
   6. PacGAN: The power of two samples in generative adversarial networks [[Paper](https://proceedings.neurips.cc/paper/2018/file/288cc0ff022877bd3df94bc9360b9c5d-Paper.pdf)]
   7. CTAB-GAN: Effective Table Data Synthesizing [[Paper](https://arxiv.org/pdf/2102.08369.pdf)]
   8. Conditional Tabular GAN-Based Two-Stage Data Generation Scheme for Short-Term Load Forecasting [[Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9253644)]
   9. TabFairGAN: Fair Tabular Data Generation with Generative Adversarial Networks [[Paper](https://arxiv.org/pdf/2109.00666.pdf)]
   10. Conditional Wasserstein GAN-based Oversampling of Tabular Data for Imbalanced Learning [[Paper](https://arxiv.org/pdf/2008.09202.pdf)]

   ##### Differential Privacy

   1. Tensorflow Privacy [[Github](https://github.com/tensorflow/privacy)]
   2. Renyi Differential Privacy [[Paper](https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/46029.pdf)]
   3. DP-CGAN : Differentially Private Synthetic Data and Label Generation [[Paper](https://arxiv.org/pdf/2001.09700.pdf)]
   4. Differentially Private Generative Adversarial Network [[Paper](https://arxiv.org/pdf/1802.06739.pdf)] [[Github](https://github.com/illidanlab/dpgan)] Another implementation [[Github](https://github.com/civisanalytics/dpwgan)]
   5. Private Data Generation Toolbox [[Github](https://github.com/BorealisAI/private-data-generation)]
   6. autodp: Automating differential privacy computation [[Github](https://github.com/yuxiangw/autodp)]
   7. Differentially Private Synthetic Medical Data Generation using Convolutional GANs [[Paper](https://arxiv.org/pdf/2012.11774.pdf)]
   8. DTGAN: Differential Private Training for Tabular GANs [[Paper](https://arxiv.org/pdf/2107.02521.pdf)]
   9. DIFFERENTIALLY PRIVATE SYNTHETIC DATA: APPLIED EVALUATIONS AND ENHANCEMENTS [[Paper](https://arxiv.org/pdf/2011.05537.pdf)]
   10. FFPDG: FAST, FAIR AND PRIVATE DATA GENERATION [[Paper](https://sdg-quality-privacy-bias.github.io/papers/SDG_paper_19.pdf)]

##### Others

   1. EvoGen: a Generator for Synthetic Versioned RDF [[Paper](http://ceur-ws.org/Vol-1558/paper9.pdf)]
   2. Generation and evaluation of synthetic patient data [[Paper](https://bmcmedresmethodol.biomedcentral.com/track/pdf/10.1186/s12874-020-00977-1.pdf)]
   3. Fake It Till You Make It: Guidelines for Effective Synthetic Data Generation [[Paper](https://www.mdpi.com/2076-3417/11/5/2158)]
   4. Generating and evaluating cross-sectional synthetic electronic healthcare data: Preserving data utility and patient privacy [[Paper](https://onlinelibrary.wiley.com/doi/epdf/10.1111/coin.12427)]
   5. Synthetic data for open and reproducible methodological research in social sciences and official statistics [[Paper](https://link.springer.com/article/10.1007/s11943-017-0214-8#Sec2)]
   6. A Study of the Impact of Synthetic Data Generation Techniques on Data Utility using the 1991 UK Samples of Anonymised Records [[Paper](https://unece.org/fileadmin/DAM/stats/documents/ece/ces/ge.46/2017/4_utility_paper.pdf)]
