Costruire Siti Web -- 2

Our site helps electronic music labels and subaggregators deliver their music to multiple DSPs. Furthermore we offer our clients tools for the management and promotion of their catalogue as well as in-depth sales reporting. Anyway, while we keep working on the current site we have started its complex porting to D7. One of the first major obstacles we are facing is the reduction of content duplication in the DB, specifically of nodes of type "release" and "track".

The catalogue of a label is comprised of releases (albums) and tracks (songs): every release has one or more tracks and every track can be present in one or more releases. Ideally each track must be unique in

the system, but unfortunately some tracks have been inserted more than once and we need to be able to easily detect those duplications and correct the problem. One way to tell if a track has duplicates is to match the title, the remix title (that is the version of the track, "Original mix" is the main version) and the artists: if two tracks have the same title, remix title and artists at 99.9% they are the same track. Obviously having a "looks like" feature with partial matches would be fantastic.

Let's have a look at two tracks that are almost identical:

--------------- TRACK A ---------------

title: Enigma Feat. Antisson

remix title: Original Vocal Mix

artists: Geo Angelo

ISRC: ITH641016999

used by release: Enigma EP - Supermarket Unlimited

--------------- TRACK B ---------------

title: Enigma Feat. Antisson

remix title: Original Vocal Mix

artists: Geo Angelo

ISRC: ITH641046297

used by release: Two Years of Supermarket - Supermarket

As you see the tracks are identical save for the ISRC code (that is a unique code). This is a case of simple substitution. Given that the track A is the correct one, the release containing the track B (Two Years of Supermarket - Supermarket) - and optionally all the releases that were to contain track B - must reference the track A instead. When all the new references are made the track B can be safely deleted. Another example, this time is a merge:

--------------- TRACK A ---------------

title: Ready to Go

remix title: Original mix

artists: Dubzap

ISRC: ITH641480849

used by release: Compressor - Musichasm Records

--------------- TRACK B ---------------

title: Ready to GO

remix title: Original mix

artists: Dubzap

ISRC: ITH641480784

duration: 00:04:27

BPM: 140

The track A is still the correct one, but the track B contains a few more information that we'd like to populate the track A with. Moreover, the track B is not used by any release. This is another operation: merge the missing information from B to A and just delete track B.

Drupal

Nº del proyecto: #6533772

Sobre el proyecto