Configuring duplicate_action

[nsteins]

I am setting up some automated imports on a collection of music downloaded from bandcamp, and there are occasional duplicates because of the way the downloader is configured. I would like to skip imports of duplicates by setting the duplicate_action tag, but is there a way to configure how duplicates are determined? The default behavior of beet duplicates relies on musicbrainz ids which does not work here, since most tracks are not on musicbrainz. I have been able to effectively find duplicates with beet duplicates -k title -k albumartist -k album however. Is there a way to make this the default method of determining duplicates for the duplicate_action tag?

[sampsyo]

Duplicate detection in the importer is, confusingly, not related at all to the duplicates plugin (and the beet duplicates command it provides). While it's only in git, #4199 added a configuration option that controls which metadata is used to detect duplicates in the importer.

[nsteins]

Thank you, this is very helpful. Curious, prior to the addition of the configuration option, what keys were used as the default for detecting duplicates?

[sampsyo]

The same as the current defaults: albumartist and album for albums and artist and title for items.