Using Public Web Services to get Information about Music Artists, Albums and Tracks

• Download Music Explorer Exe - 112.1 KB
• Download Music Explorer VS 2005 Solution - 229 KB

Introduction

Music is a big part of our lives and there is certainly no shortage of it. Hundreds of new music releases occur every day. Never in the history of civilization has so much music been produced, bought, sold and enjoyed. With all that music out there, it can be challenging to keep track of it. I use a number of online services to discover new artists and listen to samples of their music. When I find interesting artists, I like to look up their discographies and download the lyrics of their best songs. I do so with a small Windows program called Music Explorer that I wrote some time ago. Given an artist name, Music Explorer shows you a list of all albums released. For each album, you can get detailed information. For each track on a given album, you can even get the lyrics. The only thing Music Explorer doesn't do is play music, because I didn't want to get embroiled in music download and copyright issues. The figure above shows the user interface of Music Explorer.

Music Explorer runs on Windows XP and Vista. Using Music Explorer, you enter the name of an artist or group and click the Search button. Within a couple of seconds, you get a list of albums released. By default, the list is sorted by descending year, but can also be sorted by title (by clicking the Title column header). When you select an album, the right pane shows information about the album and its tracks. The album’s cover art is also shown. If you select a track, the bottom right pane shows the song’s lyrics.

Enter any artist and you can see all the lyrics of his/her songs. Pretty awesome, but Music Explorer is really not that sophisticated. It simply leverages data exposed by a small collection of public web services, showing the data in a (hopefully) intuitive way. Music Explorer is a pretty simple .Net Windows Forms application, written in C#. The application requires no installation procedure. Just copy it into a folder and run it. It doesn’t make of any third-party DLLs and will build in Visual Studio 2005 or later.

Web Services Used

Let’s look in more detail at the web services called by Music Explorer, since those provide the real added-value of the application. I originally wrote Music Explorer in 2005. Back then, the landscape of music-related web services was relatively small and immature. Amazon had a service called E-Commerce Service (ECS) 3.0 that, given an artist, would return all the albums and tracks. ECS didn’t return all the track information I wanted, so I had to call a second web service, MusicBrainz.org, an open source project created by Robert Kaye. For the lyrics, I had to use a third service, which returned lyrics as HTML that I had to scrub before showing to the user. The scrubbing process was somewhat clunky, because occasionally the service would return weird characters in the HTML and mess up the cleansed text.

Fast forward to 2009. Amazon ECS 3.0 is no longer in service. The current version is 4.0 and the name has been changed to Amazon Web Services (AWS). The new service has been heavily refactored, to reduce the number of web methods and provide better control over the amount and type of data returned. When ECS 3.0 was turned off, Music Explorer stopped working, so I set it aside with the intention of updating it at some point. That was a while ago, but I finally found some time to update the project.

When I looked at the current AWS 4.0, I discovered that Amazon has a LOT of data that can be returned. To manage retrieval of this data, Amazon resorted to a paging scheme. To give you an idea of how much data I’m talking about, a search for The Beatles returned 118 pages of data. It takes way too long to retrieve all those pages, for the purposes of Music Explorer, and most of the data is related to weird limited editions in countries that I’m interested in. The solution was to leverage MusicBrainz.org again. Back in 2005 they supported a data protocol called RDF, which has fallen out of favor. Now MusicBrainz.org exposes a REST XML web service that is pretty easy to use, once you get the hang of it. MusicBrainz also gives pretty impressive performance, giving back a list of albums in a second or two.

Regarding the lyrics web service, I found a newer service called LyricWiki.org that provides exactly what I need. Lyrics are returned as straight text, so I no longer need the HTML cleansing operation. The LyricWiki database is also pretty extensive. As of this writing, LyricWiki claims to have upwards of 700,000 pages on their site and I’m inclined to believe them. I’ve poked around with LyricWiki for a while and found that the service indeed seems to be very well stocked.

In a Nutshell

Music Explorer works by calling web services and displaying the information returned. Getting album information for an artist requires the following steps:

1. Resolving the artist’s name. When the user enters an artist name and then clicks the Search button, the artist name must be resolved to an ID before any other information can be retrieved. The MusicBrainz web service is used.
2. Getting the artist’s albums. The artist’s ID is used to retrieve a list of albums. The MusicBrainz service is used again.
3. Getting an album’s details. The album IDs (there are two of them) are used to retrieve details, which include things like the Label (e.g. Sony), genre (e.g. Heavy Metal), album cover art and the track listing. The MusicBrainz and Amazon services are used.
4. Getting a track’s lyrics. The LyricWiki service is used.

The following sequence diagram shows the details of resolving an artist name to an ID.

Figure 1 – Resolving an artist name to an ID.

Given an artist name, MusicBrainz returns a list of IDs for artists with similar names. Music Explorer looks for the ID of the artist with the exact name entered. The following sequence diagram shows the details of getting a list of albums for a given artist:

Figure 2 - Getting a list of albums for a given artist.

Each album returned by MusicBrainz includes two IDs: the MusicBrainz ID and the Amazon Standard Identification Number (ASIN). The latter ID is important, because it allows us to access the Amazon Web Service to get any additional information we might need that MusicBrainz doesn’t provide.

When the user selects an album, Music Explorer first calls Amazon, using the album ASIN, to get album details such as the genre, the number of discs in the album and the URL of the album cover art. Music Explorer then calls MusicBrainz to get the album’s track listing. The following diagram shows the details.

Figure 3 - Getting the details for an album.

When the user selects a track, Music Explorer makes a quick call to LyricWiki to retrieve the lyrics, as described in the next figure. The only parameters LyricWiki needs are the artist name and the song title.

Figure 4 - Getting the lyrics for a song.

LyricWiki returns lyrics in plain text, so you don’t have to jump through hoops to clean it up for presentation. Sweet and simple.

The Album and Track Classes

Music Explorer uses an Album class to describe music albums. An Album represents anything that stores tracks, such as a vinyl record, a cassette or a CD. The class is shown in the following listing:

public class Album
{
  public string Id;<span>       </span>// MusicBrainz ID
  public string Asin;<span>     </span>// Amazon Standard Identification Number
  public string Title;<span>    </span>// the title of the album
  public DateTime Date;<span>   </span>// the release date
  public string ImageUrl; // the URL of the cover art picture
  public string Label;<span>    </span>// e.g. Sony, Warner
  public int Discs;<span>       </span>// number of discs in album
  public string Genre;<span>    </span>// e.g. Heavy Metal
  public string Artist;<span>   </span>// name of the artist or group
  public List<Track> Tracks = new List<Track>();
}

Listing 1 - The Album class.

The Tracks field contains a list of Track objects. The Track class looks like this:

public class Track
{
  public string Id;<span>         </span>// MusicBrainz ID
  public TimeSpan duration; // length of track
  public string Title;<span>      </span>// name of song
}

Listing 2 - The Track class.

Resolving an Artist Name to an ID

Let’s look at the details. The search process starts by entering an artist name at the top of the Music Explorer screen. The name is used in a call to the MusicBrainz service, with query parameters indicating the name of the artist. The result is XML containing a list of all the artists with names similar to the one entered. For example, to search for artists with names like Madonna, Music Explorer would use the following URL:

http://musicbrainz.org/ws/1/artist/?query=Madonna&type=xml

You can enter this URL in a browser, to experiment with the MusicBrainz service. The result is a list of all the matching artists. Each artist is tagged with an id attribute. This value is important, because it must be used in subsequent searches for the artist’s albums. For the Madonna search, MusicBrainz returns the following results:

<metadata xmlns="<a href=""http://musicbrainz.org/ns/mmd-1.0"">http://musicbrainz.org/ns/mmd-1.0</a>#" 
          xmlns:ext="<a href=""http://musicbrainz.org/ns/ext-1.0"">http://musicbrainz.org/ns/ext-1.0</a>#">
    <artist-list count="11" offset="0">
        <artist id="79239441-bfd5-4981-a70c-55c3f15c1287" type="Person" ext:score="100">
            <name>Madonna</name> 
            <sort-name>Madonna</sort-name> 
            <life-span begin="1958-08-16" /> 
        </artist>
        <artist id="71851a56-9a92-422f-9ce2-c16ed0dc6938" ext:score="80">
            <name>Madonna Hip Hop Massaker</name> 
            <sort-name>Madonna Hip Hop Massaker</sort-name> 
        </artist>
        <artist id="b3d01315-d52a-4f3a-908b-0618315c1ef2" type="Person" ext:score="80">
            <name>Madonna Louise Veronica Ciccone</name> 
            <sort-name>Ciccone, Madonna Louise Veronica</sort-name> 
            <life-span begin="1958-08-16" /> 
        </artist>
        …
        <artist id="6cc0b2d5-e6b5-46c4-82c2-d4b19e721bf8" ext:score="24">
            <name>?????? ?????? ? ???? ???lt;/name> 
            <sort-name>Arnita, Madonna and Joseph Saqr</sort-name> 
        </artist>
    </artist-list>
</metadata>

Listing 3 - The artist response XML from MusicBrainz.

I omitted most of the <artist> results, for clarity. Note that non-Latin character can appear in the results. Music Explorer uses the following XPATH code to extract the information for the artist with the exact name entered.

public static string FindArtistId(string name)
{
  string uri = string.Format(<a href=""http://musicbrainz.org/ws/1/artist/?query=%7b0%7d&type=xml""><span>http://musicbrainz.org/ws/1/artist/?query={0}&type=xml</span></a>", 
                             name);
  XPathDocument doc = new XPathDocument(uri);
  XPathNavigator nav = doc.CreateNavigator();
  XmlNamespaceManager nsmgr = new XmlNamespaceManager(nav.NameTable);
  nsmgr.AddNamespace("mb", "http://musicbrainz.org/ns/mmd-1.0#");
  string xpath = string.Format("//mb:artist[mb:name=\"{0}\"]", name);
  XPathNodeIterator ni = nav.Select(xpath, nsmgr);
  if (!ni.MoveNext()) return null;
  XPathNavigator current = ni.Current;
  return current.GetAttribute("id", nsmgr.DefaultNamespace);
}

Listing 4 - Extracting the artist's id from the response XML.

<metadata xmlns="<a href=""http://musicbrainz.org/ns/mmd-1.0"">http://musicbrainz.org/ns/mmd-1.0</a>#">
    <artist id="79239441-bfd5-4981-a70c-55c3f15c1287" type="Person">
        <name>Madonna</name> 
        <sort-name>Madonna</sort-name> 
        <life-span begin="1958-08-16" /> 
        <release-list>
            <release id="5d9a5a55-e75f-4282-906b-1141f94f1885" type="Album Official">
              <title>Madonna</title> 
              <text-representation language="ENG" script="Latn" /> 
              <asin>B000002KZ3</asin> 
              <release-event-list>
                <event date="1983-07-26" country="US" catalog-number="1-23867"
                       barcode="075992386715" format="Vinyl" /> 
                <event date="1983-07-26" country="US" catalog-number="9 23867-4" 
                        barcode="075992386746" format="Cassette" /> 
                <event date="1983-07-26" country="US" catalog-number="9 23867-2" 
                       barcode="075992386722" format="CD" /> 
                <event date="1983-07-27" country="JP" catalog-number="PKF-5423" 
                       format="Cassette" /> 
              </release-event-list>
            </release>
            ...
            ...
      </release-list>
    </artist>
</metadata>

<metadata xmlns=http://musicbrainz.org/ns/mmd-1.0#
          xmlns:ext="<a href=""http://musicbrainz.org/ns/ext-1.0"">http://musicbrainz.org/ns/ext-1.0</a>#">
  <release id="5d9a5a55-e75f-4282-906b-1141f94f1885" type="Album Official">
    <title>Madonna</title> 
      <text-representation language="ENG" script="Latn" /> 
      <asin>B000002KZ3</asin> 
      <track-list>
         <track id="9a501c4e-c7cb-40f8-ae2d-6cd92807f9a8">
           <title>Lucky Star</title> 
           <duration>337106</duration> 
         </track>
         <track id="516012bf-ff23-4fb9-ae18-d70f35fa8603">
           <title>Borderline</title> 
           <duration>321893</duration> 
         </track>
         ...
      </track-list>
  </release>
</metadata>

ItemLookup search = new ItemLookup();
search.AssociateTag = Properties.Settings.Default.AmazonAwsAccountNumber;
search.AWSAccessKeyId = Properties.Settings.Default.AmazonAwsAccessKeyId;
ItemLookupRequest request = new ItemLookupRequest();
request.ItemId = new string[] { album.Asin };
request.IdType = ItemLookupRequestIdType.ASIN;
request.ResponseGroup = new string[] {"Images","ItemAttributes","BrowseNodes"};
search.Request = new ItemLookupRequest[] { request };
ItemLookupResponse response = amazonWebService.ItemLookup(search);

AmazonAws.Items items = response.Items[0];
if (items.Item == null) return;
if (items.Item.Length == 0) return;
AmazonAws.Item item = items.Item[0];
if (item == null) return;
if (item.MediumImage == null)
  album.ImageUrl = null;
else
  album.ImageUrl = item.MediumImage.URL;
if (item.ItemAttributes != null)
  album.Label = item.ItemAttributes.Label;
if (item.ItemAttributes.NumberOfDiscs != null)
  album.Discs = int.Parse(item.ItemAttributes.NumberOfDiscs);
album.Genre = GetGenre(item);

static string GetGenre(AmazonAws.Item item)
{
  if (item.BrowseNodes == null) return null;
  string genre = string.Empty;
  foreach (BrowseNode node in item.BrowseNodes.BrowseNode)
  {
    if (node.Name.ToLower() == "general")
    if (node.Ancestors != null)
    return node.Ancestors[0].Name;
  }
  return "<unspecified>";
}

void ShowCoverArt()
{
  if (album.ImageUrl == null)
  {
    pictureBoxAlbumCover.Image = notAvailable;
    return;
  }
  WebClient client = new WebClient();
  client.Headers["User-Agent"] = "Mozilla/4.0";
  byte[] bytes = client.DownloadData(album.ImageUrl.ToString());
  MemoryStream stream = new MemoryStream(bytes);
  pictureBoxAlbumCover.Image = System.Drawing.Image.FromStream(stream);
}