querio 0.0.2

A python library for quickly estimating database results

Quer.io

Documentation links

See Documentation for Documentation

See Usage guide for a basic rundown on how to use Quer.io

See Database Schema 1 for single table sample database schema

See Database Schema 2 for normalized sample database schema

See ML documentation for documentation on the machine learning model

Project description

This project is built to the specifications and requirements provided by Prof. Michael Mathioudakis and is a course work project for course TKT20007 Software Engineering Lab at the University of Helsinki, department of Computer Science.

The aim of this project is to build an Approximate Query Processing (AQP) engine -- i.e., a software layer on top of a relational database, that allows us to obtain fast, approximate answers to aggregate queries, with the help of Machine Learning models.

Chosen implementation is a Python library that can be used with multiple different database systems. Machine learning components are built using Scikit Learn.

Installation

This project is done with Python 3.6

See Database Installation guide for information how to install the sample databases this application was tested on.

See Application Installation guide for information how to install the application and all its dependencies.

Optional installation

See Querio Scheduler for how to install and use a scheduler for periodical model retraining.

Tests

Currently the project contains tests that are done using the unittest library. Tests can be run with the following command from the project root

python3 -m unittest discover

This command will find every test from the project and run it. If you want to run an individual test script it can be done with the following command

python3 -m unittest [path to file]

Contributors

• Dennis Ahlfors
• Joonas J
• Kim Toivonen
• Mauri Frestadius
• Ossi Räisä
• Petja Valkama