MusicBottle (or "the Bottle") is a project with the intend of presenting MusicBrainz data in a way more suitable for presentation and browsing where musicbrainz.org is mostly for editing purposes. It is led by Freso and caller#6 with contributions from Lordsputnik and a bunch of other the MusicBrainz people.
Discussion will happen probably mostly on IRC in the room #musicbottle. Feel free to come and hang out. We will probably also utilise the mb-devel mailing list for getting feedback on some design decisions (e.g., input on how to implement caching).
- Source code on GitHub.
- Code running on the MusicBrainz sandbox server. (Not guaranteed to be running.)
- Continuous integration (Jenkins) (Test results from whenever new code is pushed to GitHub.)
- @musicbottle_dev (Twitter)
- @musicbottledev (Identi.ca)
Purpoise, the Mission Statement
- Showcase the MusicBrainz data in a prettier way than what can reasonably be achieved on musicbrainz.org.
- Show what can be done using the data fetched from MusicBrainz' Web Service.
- Don't hardcode any information. While it may be tempting to mark that Mozart is a composer in the code, we want the site to display the what can be retrieved using only MusicBrainz. Retrieving information from sources linked to from MB data is fine, but hardcoding such info is not.
- Use the URL scheme of musicbrainz.org as long as at makes sense. This should simplify a lot of things, and make inter-site URL mapping quite easy. So e.g. http://musicbottle.net/artist/3c25958a-2bff-4381-8eb4-7dbe84c3e75e will match https://musicbrainz.org/artist/3c25958a-2bff-4381-8eb4-7dbe84c3e75e
As Freso likes writing Python and caller#6 wants to learn the language, this quickly became the language of the project. Flask was initially recommended by ianweller and subsequently endorsed by ruaok, so that is the framework we're currently looking into. We will most likely aim for getting the code running on an nginx server.
We're also going to need some kind of cache, to not hit MusicBrainz (or other sites) over and over again. We're looking into MongoDB and possibly other NoSQL databases for now.
The main codebase is GPL v2+ and any libraries developed to keep sane abstractions will be LGPL'd. (Any Flask-specific extensions will be BSD/MIT-style licensed.) The code is available on GitHub.
Libraries, APIs, etc.
- Web services:
- MusicBrainz: XML / JSON
- MediaWiki API
- YouTube API
- Twitter Developers
- Facebook Developers
- Bandcamp API (Terms)
- Cover Art Archive
- Fanart.tv Music API
- MySpace Developer Platform
- iTunes Search API
- Discogs API
- Flickr Machine Tag
- Last.fm API (provides lookup of artist by MBID and images for artists)
- WorldCat Affiliate Tools
- Python libraries:
- Coding style (Pocoo Styleguide)
- @seealso Code Like a Pythonista: Idiomatic Python for a tour through PEP 8 and various Python "tricks".
- See JSON output "pretty-printed" in your browser:
- Examples of Flask usage
How and what to pull in and how to present it?
Twitter, YouTube, Facebook, ... activity streams.
Images from fanart.tv
- Music Web Navigator - uses MusicBrainz and Last.FM data.
- "Moto Boy (or, how to build a beautiful and functional artist website)"
- BBC Music artist pages (e.g. S Club (7))
- Beardscratchers Compendium of Music