mirror of
https://github.com/Alfresco/alfresco-community-repo.git
synced 2025-10-08 14:51:49 +00:00
125603 rmunteanu: Merged 5.1.1 (5.1.1) to 5.1.N (5.1.2) 125484 slanglois: MNT-16155 Update source headers - remove old Copyrights from Java and JSP dource files git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/BRANCHES/DEV/5.2.N/root@125781 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
165 lines
4.9 KiB
Java
165 lines
4.9 KiB
Java
package org.alfresco.repo.content.metadata;
|
|
|
|
import java.io.Serializable;
|
|
import java.util.ArrayList;
|
|
import java.util.Calendar;
|
|
import java.util.Date;
|
|
import java.util.Map;
|
|
|
|
import org.alfresco.repo.content.MimetypeMap;
|
|
import org.apache.tika.config.TikaConfig;
|
|
import org.apache.tika.metadata.Metadata;
|
|
import org.apache.tika.metadata.XMPDM;
|
|
import org.apache.tika.parser.CompositeParser;
|
|
import org.apache.tika.parser.Parser;
|
|
import org.apache.tika.parser.mp4.MP4Parser;
|
|
import org.gagravarr.tika.FlacParser;
|
|
import org.gagravarr.tika.VorbisParser;
|
|
|
|
/**
|
|
* A Metadata Extractor which makes use of the Apache
|
|
* Tika Audio Parsers to extract metadata from your
|
|
* media files.
|
|
* For backwards compatibility reasons, this doesn't
|
|
* handle the MP3 format, which has its own dedicated
|
|
* extractor in {@link MP3MetadataExtracter}
|
|
|
|
* <pre>
|
|
* <b>author:</b> -- cm:author
|
|
* <b>title:</b> -- cm:title
|
|
* <b>created:</b> -- cm:created
|
|
* <b>xmpDM:artist</b> -- audio:artist
|
|
* <b>xmpDM:composer</b> -- audio:composer
|
|
* <b>xmpDM:engineer</b> -- audio:engineer
|
|
* <b>xmpDM:genre</b> -- audio:genre
|
|
* <b>xmpDM:trackNumber</b> -- audio:trackNumber
|
|
* <b>xmpDM:releaseDate</b> -- audio:releaseDate
|
|
* </pre>
|
|
*
|
|
* @since 4.0
|
|
* @author Nick Burch
|
|
*/
|
|
public class TikaAudioMetadataExtracter extends TikaPoweredMetadataExtracter
|
|
{
|
|
protected static final String KEY_LYRICS = "lyrics";
|
|
|
|
// The Audio related parsers we use
|
|
private static Parser[] parsers = new Parser[] {
|
|
new VorbisParser(),
|
|
new FlacParser(),
|
|
new MP4Parser()
|
|
};
|
|
// The explicit mimetypes we support (plus any others from the parsers)
|
|
public static ArrayList<String> SUPPORTED_MIMETYPES = buildSupportedMimetypes(
|
|
new String[] {
|
|
MimetypeMap.MIMETYPE_VORBIS, MimetypeMap.MIMETYPE_FLAC,
|
|
MimetypeMap.MIMETYPE_AUDIO_MP4,
|
|
}, parsers
|
|
);
|
|
|
|
protected TikaConfig tikaConfig;
|
|
public void setTikaConfig(TikaConfig tikaConfig)
|
|
{
|
|
this.tikaConfig = tikaConfig;
|
|
}
|
|
|
|
public TikaAudioMetadataExtracter()
|
|
{
|
|
this(SUPPORTED_MIMETYPES);
|
|
}
|
|
public TikaAudioMetadataExtracter(ArrayList<String> supportedMimeTypes)
|
|
{
|
|
super(supportedMimeTypes);
|
|
}
|
|
|
|
@Override
|
|
protected Parser getParser()
|
|
{
|
|
return new CompositeParser(
|
|
tikaConfig.getMediaTypeRegistry(), parsers
|
|
);
|
|
}
|
|
|
|
@Override
|
|
protected Map<String, Serializable> extractSpecific(Metadata metadata,
|
|
Map<String, Serializable> properties, Map<String,String> headers)
|
|
{
|
|
// Most things can go with the default Tika -> Alfresco Mapping
|
|
// Handle the few special cases here
|
|
|
|
// The description is special
|
|
putRawValue(KEY_DESCRIPTION, generateDescription(metadata), properties);
|
|
|
|
// The release date can be fiddly
|
|
Date releaseDate = generateReleaseDate(metadata);
|
|
putRawValue(KEY_CREATED, releaseDate, properties);
|
|
putRawValue(XMPDM.RELEASE_DATE.getName(), releaseDate, properties);
|
|
|
|
// TODO Get the Lyrics from the content
|
|
//putRawValue(KEY_LYRICS, getLyrics(), properties);
|
|
|
|
// All done
|
|
return properties;
|
|
}
|
|
|
|
/**
|
|
* Generates the release date
|
|
*/
|
|
private Date generateReleaseDate(Metadata metadata)
|
|
{
|
|
String date = metadata.get(XMPDM.RELEASE_DATE);
|
|
if(date == null || date.length() == 0)
|
|
{
|
|
return null;
|
|
}
|
|
|
|
// Is it just a year?
|
|
if(date.matches("\\d\\d\\d\\d"))
|
|
{
|
|
// Just a year, we need a full date
|
|
// Go for the 1st of the 1st
|
|
Calendar c = Calendar.getInstance();
|
|
c.set(
|
|
Integer.parseInt(date), Calendar.JANUARY, 1,
|
|
0, 0, 0
|
|
);
|
|
c.set(Calendar.MILLISECOND, 0);
|
|
return c.getTime();
|
|
}
|
|
|
|
// Treat as a normal date
|
|
return makeDate(date);
|
|
}
|
|
|
|
/**
|
|
* Generate the description
|
|
*
|
|
* @param metadata the metadata extracted from the file
|
|
* @return the description
|
|
*/
|
|
@SuppressWarnings("deprecation")
|
|
private String generateDescription(Metadata metadata)
|
|
{
|
|
StringBuilder result = new StringBuilder();
|
|
if (metadata.get(Metadata.TITLE) != null)
|
|
{
|
|
result.append(metadata.get(Metadata.TITLE));
|
|
if (metadata.get(XMPDM.ALBUM) != null)
|
|
{
|
|
result
|
|
.append(" - ")
|
|
.append(metadata.get(XMPDM.ALBUM));
|
|
}
|
|
if (metadata.get(XMPDM.ARTIST) != null)
|
|
{
|
|
result
|
|
.append(" (")
|
|
.append(metadata.get(XMPDM.ARTIST))
|
|
.append(")");
|
|
}
|
|
}
|
|
|
|
return result.toString();
|
|
}
|
|
}
|