New features in AcousticBrainz

We’ve been making a number of changes to AcousticBrainz over the last few months, but haven’t been publicly announcing the changes. This post (hopefully a thread in the future?) is to remedy this and keep everyone up-to-date on what we’ve been doing.

4 June

  • Add a submission offset to the database (AB-406). This was a backend change that optimised our database and helped us reduce some of the load on the AcousticBrainz server, which was causing slowness issues. Thanks to our Summer of Code student @aidanlw17 for this!

1 July

  • Map model classes to more meaningful labels (AB-91). Many of our high level models have short cryptic names. We made some changes to show easier to understand names on the recording detail pages and in the API if you add a parameter. @pulkit6559 did much of the initial work on this

  • Include dataset and class descriptions in CSV for import/export (AB-387). We had the ability to export the contents of datasets, but this didn’t include all of the information that people were able to add (dataset description and class descriptions). Now, exporting a dataset to CSV will include this information, and importing a CSV with this information will add it to the dataset. Thanks to @aidanlw17

  • Fix errors when uuids have upper-case characters (AB-411). It turns out that we had a long-standing bug where requesting a MBID with upper-case letters wouldn’t work, this works now.

  • Improve storage and retrieval of statistics (AB-385). This backend change allowed us to improve the way that we collect and show statistics on the statistics page. Thanks to @pulkit6559 for this

31 July

Any questions or bug reports, we’re always here, and on the metabrainz tickets platform. Enjoy!