graphomaly 0.1

software package for anomaly detection in graphs modeling financial transactions

Graphomaly

The Anti-Money Laundering (AML) tool. Find abnormal data in graph and network structures.

Official package documentation here.

Installation and setup

Install via pip:

pip install https://gitlab.com/unibuc/graphomaly/graphomaly

A package will be uploaded to PyPI soon.

Usage

The package follows the sklearn API and can be included in your projects via

from graphomaly.estimator import GraphomalyEstimator

which will provide you with a standard scikit-learn estimator that you can use in your pipeline.

For configuration and tweaks please consult the YAML file for now until documentation matures.

Development and testing

First clone the repository and change directory to the root of your fresh checkout.

0. Install Prerequisites

Install PyPA’s build:

python3 -m pip install --upgrade build

1. Build

Inside the Graphomaly directory

python -m build

2. Virtual Environment

Create a virtual environment with Python:

python -m venv venv

Activate the environment:

source venv/bin/activate

For Windows execute instead:

venv\Scripts\activate

3. Install

Inside the virutal environment execute:

pip install dist/graphomaly-*.whl

Running unit tests

First create the results directory:

mkdir -p tests/results/synthetic

Run the initial test on synthetic data to make sure things installed ok:

cd tests && python test_synthetic

Then run the other unit tests by hand as above or via pytest:

pytest  # add -v for verbose, add -s to print(...) to console from tests