HoloClean: A Machine Learning System for Data Enrichment

HoloClean is built on top of PyTorch and Postgres.

HoloClean is a statistical inference engine to impute, clean, and enrich data. As a weakly supervised machine learning system, HoloClean leverages available quality rules, value correlations, reference data, and multiple other signals to build a probabilistic model that accurately captures the data generation process, and uses the model in a variety of data curation tasks. HoloClean allows data practitioners and scientists to save the enormous time they spend in building piecemeal cleaning solutions, and instead, effectively communicate their domain knowledge in a declarative way to enable accurate analytics, predictions, and insights form noisy, incomplete, and erroneous data.

Installation

1. Install Postgres

Ubuntu

Install Postgres by running

$ apt-get install postgresql postgresql-contrib

MacOS

Installation instructions can be found at https://www.postgresql.org/download/macosx/.

2. Setup Postgres for HoloClean

To start the Postgres console from your terminal

$ psql --user <username>    # you can omit --user <username> to use current user

We then create a database holo and user holo (default settings for HoloClean)

CREATE DATABASE holo;
CREATE USER holocleanuser;
ALTER USER holo WITH PASSWORD 'abcd1234';
GRANT ALL PRIVILEGES ON DATABASE holo TO holocleanuser;
\c holo
ALTER SCHEMA public OWNER TO holo;

In general, to connect to the holo database from the Postgres psql console

\c holo

HoloClean currently populates the database holo with auxiliary and meta tables. To clear the database simply connect as a root user or as holocleanuser and run

DROP DATABASE holo;
CREATE DATABASE holo;

3. Set up HoloClean

Virtual Environment

Option 1: Set up a conda Virtual Environment

Install Conda using one of the following methods

Ubuntu

For 32-bit machines run

$ wget https://repo.continuum.io/archive/Anaconda-2.3.0-Linux-x86.sh
$ sh Anaconda-2.3.0-Linux-x86.sh

For 64-bit machines run

$ wget https://repo.continuum.io/archive/Anaconda-2.3.0-Linux-x86_64.sh
$ sh Anaconda-2.3.0-Linux-x86_64.sh

MacOS

Follow instructions here to install Anaconda (NOT miniconda).

Create a conda environment

Create a Python 2.7 conda environment by running

$ conda create -n holo_env python=2.7

Upon starting/restarting your terminal session, you will need to activate your conda environment by running

$ source activate holo_env

NOTE: ensure your environment is activated throughout the installation process.

Option 2: Set up a virtual environment using pip and Virtualenv

If you are familiar with Virtualenv, create a new Python 2.7 environment with your preferred Virtualenv wrapper, for example:

• virtualenvwrapper (Bourne-shells)
• virtualfish (fish-shell)

Install Virtualenv

Either follow instructions here or install via pip

$ pip install virtualenv

Create a Virtualenv environment

Create a new directory for a Python 2.7 virtualenv environment

$ mkdir -p holo_env
$ virtualenv --python=python holo_env

where python is a valid reference to a python executable.

Activate the environment

$ source bin/activate

NOTE: ensure your environment is activated throughout the installation process.

Install the requirements of HoloClean

In the project root directory, run the following to install the required packages. Note that this commands installs the packages within the activated virtual environment.

$ pip install -r requirements.txt

Note about MacOS

If you are on MacOS, you may need to install XCode developer tools using the command xcode-select --install.

Usage

See the code in tests/test_holoclean.py for a documented usage of HoloClean.

In order to run the test script, run the following:

$ cd tests
$ ./start_test.sh

The script sets up the python path environment for running holoclean.