The Django Elvis Web Application
Welcome to the ELVIS Django database of symbolic music. This application is intended to allow for uploading, searching, and downloading symbolic music files for use in music analysis frameworks, such as the VIS Web App.
Models in the Django app include composers, movements, collections, and many more. Here are some comments on what some of these models are meant for:
- The attachment model contains files associated with each piece or movement of music. A movement may have multiple attachments if its been transcribed by different people.
- The collection model is to be used for keeping or sharing sets of pieces/movements that are of particular interest. These collections used to be called corpora but were since renamed.
- The download model is a "shopping cart" of attachments associated with each user. A user would query the database for pieces/movements, add relevant attachments to their downloads "cart", and "checkout" the files by downloading them when needed. This was vaguely inspired by online shopping sites.
- The genre, instrumentation, language, location, and source models will enhance the old tagging system. Pieces/movements will be tagged with particular instances of these models for easy searching and to structure the metadata associated with the musical content.
We plug the Django app into Apache Solr to speed up text-based searches and to allow for queries with complex logical structure. It lets us use a faceted search UI as well. We also use Celery with Rabbitmq as broker to coordinate downloading. At checkout, attachments in a cart is zipped into a temporary file and served via a call to a Celery job. A Celery beat job periodically cleans out the zipped files after they've been downloaded.
A deployment guide is including in the docs of this repository with details as to how we deployed the database on Fedora 19.
See GitHub issues.