Skip to main content

A collection of tools that simplify the downloading and handling of datasets used for Optical Music Recognition (OMR).

Project description

Tools for working with Optical Music Recognition datasets

Build Status codecov PyPI version Documentation Status GitHub license

A collection of tools that simplify the downloading and handling of datasets used for Optical Music Recognition (OMR). These tools are available as Python package omrdatasettools on PyPi.

They simplify the most common tasks such as downloading and extracting a dataset, generating images from textual representations or visualizing those datasets.

Development setup

Create virtual environment running

python3.11 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Running tests

Changelog

1.4.0

Adding several new datasets, upgrading to Python 3.11, and refactoring downloader to use Pathlib and support tar-gz.

1.3.1

Bugfixing an error that generates incorrect MUSCIMA++ image masks because images were incorrectly associated with each other.

1.3.0

Added download capabilities of DeepScores V1 with extended vocabulary and opened the downloader, so you can download custom datasets, as well as utilize other methods from it that were previously private.

1.2.2

Fixed incorrect import statement in __init__.py

1.2.1

Fixed dependency problem during setup.py that prevented the package from being installed if the dependent libraries are not yet installed (which defeats the purpose of declaring dependencies in setup.py). Changing to semantic versioning with three numbers. 1.2 is now considered 1.2.0.

1.2

Attempting to declare dependencies in setup.py properly

1.1

Updated MuscimaPlusPlusSymbolImageGenerator to work with MUSCIMA++ 2.0. Added quality-of-life improvement suggested by @yvan674 to make importing common classes such as the downloader easier.

1.0

Dramatically simplified the tools for downloading datasets. Removed mostly unused code and re-organized project structure and documentation.

0.19

New Image generator that can take MUSCIMA++ v2.0 images and generate masks for instance segmentation of staffs, as well as masks for semantic segmentation for all objects.

0.18

Changing MUSCIMA++ Downloader to accept a string instead of integer for enabling future versioning of the dataset beyond integers, e.g., "2.1".

Previous releases

For information on previous releases, check out the Github Repository

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

omrdatasettools-1.4.0.tar.gz (41.5 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page