* clean release info before submitting queries

source/net/sourceforge/filebot/format/MediaBindingBean.java

@@ -2,28 +2,21 @@
 package net.sourceforge.filebot.format;
 
 
-import static java.util.Arrays.*;
-import static java.util.ResourceBundle.*;
-import static java.util.regex.Pattern.*;
 import static net.sourceforge.filebot.MediaTypes.*;
 import static net.sourceforge.filebot.format.Define.*;
 import static net.sourceforge.filebot.hash.VerificationUtilities.*;
 
 import java.io.File;
 import java.io.IOException;
-import java.nio.ByteBuffer;
-import java.nio.charset.Charset;
 import java.util.Scanner;
-import java.util.regex.Matcher;
-import java.util.regex.Pattern;
 
 import net.sf.ehcache.Cache;
 import net.sf.ehcache.CacheManager;
 import net.sf.ehcache.Element;
 import net.sourceforge.filebot.hash.HashType;
 import net.sourceforge.filebot.mediainfo.MediaInfo;
+import net.sourceforge.filebot.mediainfo.ReleaseInfo;
 import net.sourceforge.filebot.mediainfo.MediaInfo.StreamKind;
-import net.sourceforge.filebot.web.CachedResource;
 import net.sourceforge.filebot.web.Date;
 import net.sourceforge.filebot.web.Episode;
 import net.sourceforge.filebot.web.Movie;
@@ -249,21 +242,10 @@ public class MediaBindingBean {
 	public String getVideoSource() {
 		// use inferred media file
 		File inferredMediaFile = getInferredMediaFile();
-		
+		ReleaseInfo releaseInfo = new ReleaseInfo();
-		// pattern matching any video source name
-		Pattern source = compile(getBundle(getClass().getName()).getString("pattern.video.source"), CASE_INSENSITIVE);
 		
 		// look for video source patterns in media file and it's parent folder
-		String lastMatch = null;
+		return releaseInfo.getVideoSource(inferredMediaFile);
-		for (File it : asList(inferredMediaFile.getParentFile(), inferredMediaFile)) {
-			for (String part : it.getName().split("[^\\p{Alnum}]")) {
-				if (source.matcher(part).matches()) {
-					lastMatch = part;
-				}
-			}
-		}
-		
-		return lastMatch;
 	}
 	
 
@@ -271,19 +253,10 @@ public class MediaBindingBean {
 	public String getReleaseGroup() throws IOException {
 		// use inferred media file
 		File inferredMediaFile = getInferredMediaFile();
-		
+		ReleaseInfo releaseInfo = new ReleaseInfo();
-		// pattern matching any release group name enclosed in separators
-		Pattern groups = compile("(?<!\\p{Alnum})(" + releaseGroups.get() + ")(?!\\p{Alnum})", CASE_INSENSITIVE);
 		
 		// look for release group names in media file and it's parent folder
-		String lastMatch = null;
+		return releaseInfo.getReleaseGroup(inferredMediaFile);
-		for (File it : asList(inferredMediaFile.getParentFile(), inferredMediaFile)) {
-			for (Matcher matcher = groups.matcher(it.getName()); matcher.find();) {
-				lastMatch = matcher.group();
-			}
-		}
-		
-		return lastMatch;
 	}
 	
 
@@ -417,14 +390,4 @@ public class MediaBindingBean {
 		return hash;
 	}
 	
-
-	// fetch release group names online and try to update the data once per day
-	private final CachedResource<String> releaseGroups = new CachedResource<String>(getBundle(getClass().getName()).getString("url.release-groups"), 24 * 60 * 60 * 1000) {
-		
-		@Override
-		public String process(ByteBuffer data) {
-			return compile("\\s").matcher(Charset.forName("UTF-8").decode(data)).replaceAll("|");
-		}
-	};
-	
 }

source/net/sourceforge/filebot/mediainfo/ReleaseInfo.java

@@ -0,0 +1,106 @@
+
+package net.sourceforge.filebot.mediainfo;
+
+
+import static java.util.ResourceBundle.*;
+import static java.util.regex.Pattern.*;
+import static net.sourceforge.tuned.StringUtilities.*;
+
+import java.io.File;
+import java.io.IOException;
+import java.nio.ByteBuffer;
+import java.nio.charset.Charset;
+import java.util.ArrayList;
+import java.util.List;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+
+import net.sourceforge.filebot.web.CachedResource;
+
+
+public class ReleaseInfo {
+	
+	public String getVideoSource(File file) {
+		// check parent and itself for group names
+		return matchLast(getVideoSourcePattern(), file.getParent(), file.getName());
+	}
+	
+
+	public String getReleaseGroup(File file) throws IOException {
+		// check parent and itself for group names
+		return matchLast(getReleaseGroupPattern(), file.getParent(), file.getName());
+	}
+	
+
+	protected String matchLast(Pattern pattern, CharSequence... sequence) {
+		String lastMatch = null;
+		
+		for (CharSequence name : sequence) {
+			if (name == null)
+				continue;
+			
+			Matcher matcher = pattern.matcher(name);
+			while (matcher.find()) {
+				lastMatch = matcher.group();
+			}
+		}
+		
+		return lastMatch;
+	}
+	
+
+	public List<String> clean(Iterable<String> items) {
+		return clean(items, getVideoSourcePattern(), getCodecPattern());
+	}
+	
+
+	public List<String> cleanRG(Iterable<String> items) throws IOException {
+		return clean(items, getReleaseGroupPattern(), getVideoSourcePattern(), getCodecPattern());
+	}
+	
+
+	public List<String> clean(Iterable<String> items, Pattern... blacklisted) {
+		List<String> cleaned = new ArrayList<String>();
+		
+		for (String string : items) {
+			for (Pattern it : blacklisted) {
+				string = it.matcher(string).replaceAll("");
+			}
+			
+			cleaned.add(string.replaceAll("[\\p{Punct}\\p{Space}]+", " ").trim());
+		}
+		
+		return cleaned;
+	}
+	
+
+	public Pattern getCodecPattern() {
+		// pattern matching any video source name
+		String pattern = getBundle(getClass().getName()).getString("pattern.codec");
+		return compile("(?<!\\p{Alnum})(" + pattern + ")(?!\\p{Alnum})", CASE_INSENSITIVE);
+	}
+	
+
+	public Pattern getVideoSourcePattern() {
+		// pattern matching any video source name
+		String pattern = getBundle(getClass().getName()).getString("pattern.video.source");
+		return compile("(?<!\\p{Alnum})(" + pattern + ")(?!\\p{Alnum})", CASE_INSENSITIVE);
+	}
+	
+
+	public Pattern getReleaseGroupPattern() throws IOException {
+		// pattern matching any release group name enclosed in separators
+		return compile("(?<!\\p{Alnum})(" + join(releaseGroupResource.get(), "|") + ")(?!\\p{Alnum})", CASE_INSENSITIVE);
+	}
+	
+
+	// fetch release group names online and try to update the data once per day
+	protected final CachedResource<String[]> releaseGroupResource = new CachedResource<String[]>(getBundle(getClass().getName()).getString("url.release-groups"), 24 * 60 * 60 * 1000) {
+		
+		@Override
+		public String[] process(ByteBuffer data) {
+			return compile("\\s").split(Charset.forName("UTF-8").decode(data));
+		}
+	};
+	
+}

source/net/sourceforge/filebot/mediainfo/ReleaseInfo.properties

@@ -1,5 +1,8 @@
 # source names mostly copied from [http://en.wikipedia.org/wiki/Pirated_movie_release_types]
 pattern.video.source: CAMRip|CAM|TS|TELESYNC|PDVD|TS|TELESYNC|PDVD|PPV|PPVRip|Screener|SCR|SCREENER|DVDSCR|DVDSCREENER|BDSCR|R5|R5LINE|DVDRip|DVDR|TVRip|DSR|PDTV|HDTV|DVBRip|DTHRip|VODRip|VODR|BDRip|BRRip|BluRay|BDR
 
+# additional release info patterns
+pattern.codec: DivX|Xvid|AVC|x264|h264|3ivx|mpeg|mpeg4|mp3|aac|2ch|6ch|720p|1080p
+
 # group names mostly copied from [http://scenelingo.wordpress.com/list-of-scene-release-groups]
-url.release-groups: http://filebot.sourceforge.net/data/release-groups.txt
+url.release-groups: http://filebot.sourceforge.net/data/release-groups.txt

source/net/sourceforge/filebot/ui/rename/MovieHashMatcher.java

@@ -13,11 +13,11 @@ import java.io.File;
 import java.io.FileInputStream;
 import java.io.IOException;
 import java.util.ArrayList;
+import java.util.Collection;
 import java.util.Collections;
 import java.util.Comparator;
 import java.util.HashMap;
 import java.util.HashSet;
-import java.util.LinkedHashSet;
 import java.util.List;
 import java.util.Locale;
 import java.util.Map;
@@ -34,6 +34,7 @@ import javax.swing.Action;
 import javax.swing.SwingUtilities;
 
 import net.sourceforge.filebot.Analytics;
+import net.sourceforge.filebot.mediainfo.ReleaseInfo;
 import net.sourceforge.filebot.similarity.Match;
 import net.sourceforge.filebot.ui.SelectDialog;
 import net.sourceforge.filebot.web.Movie;
@@ -160,11 +161,6 @@ class MovieHashMatcher implements AutoCompleteMatcher {
 	}
 	
 
-	private String normalizeMovieName(File movie) {
-		return getName(movie).replaceAll("\\p{Punct}+", " ").trim();
-	}
-	
-
 	protected Movie grabMovieName(File movieFile, Locale locale, boolean autodetect, Movie... suggestions) throws Exception {
 		List<Movie> options = new ArrayList<Movie>();
 		
@@ -185,9 +181,12 @@ class MovieHashMatcher implements AutoCompleteMatcher {
 		}
 		
 		// search by file name or folder name
-		Set<String> searchQueries = new LinkedHashSet<String>(2);
+		Collection<String> searchQueries = new TreeSet<String>(String.CASE_INSENSITIVE_ORDER);
-		searchQueries.add(normalizeMovieName(movieFile));
+		searchQueries.add(getName(movieFile));
-		searchQueries.add(normalizeMovieName(movieFile.getParentFile()));
+		searchQueries.add(getName(movieFile.getParentFile()));
+		
+		// remove blacklisted terms
+		searchQueries = new ReleaseInfo().cleanRG(searchQueries);
 		
 		for (String query : searchQueries) {
 			if (autodetect && options.isEmpty()) {
@@ -197,7 +196,7 @@ class MovieHashMatcher implements AutoCompleteMatcher {
 		
 		// allow manual user input
 		if (options.isEmpty() || !autodetect) {
-			String suggestion = options.isEmpty() ? normalizeMovieName(movieFile) : options.get(0).getName();
+			String suggestion = options.isEmpty() ? searchQueries.iterator().next() : options.get(0).getName();
 			String input = showInputDialog(null, "Enter movie name:", suggestion);
 			
 			if (input != null) {

test/net/sourceforge/filebot/FileBotTestSuite.java

@@ -8,6 +8,7 @@ import org.junit.runners.Suite.SuiteClasses;
 
 import net.sourceforge.filebot.format.ExpressionFormatTest;
 import net.sourceforge.filebot.hash.VerificationFormatTest;
+import net.sourceforge.filebot.mediainfo.ReleaseInfoTest;
 import net.sourceforge.filebot.similarity.SimilarityTestSuite;
 import net.sourceforge.filebot.subtitle.SubtitleReaderTestSuite;
 import net.sourceforge.filebot.ui.rename.MatchModelTest;
@@ -16,7 +17,7 @@ import net.sourceforge.filebot.web.WebTestSuite;
 
 
 @RunWith(Suite.class)
-@SuiteClasses( { SimilarityTestSuite.class, WebTestSuite.class, ExpressionFormatTest.class, VerificationFormatTest.class, MatchModelTest.class, MatchSimilarityMetricTest.class, SubtitleReaderTestSuite.class })
+@SuiteClasses( { SimilarityTestSuite.class, WebTestSuite.class, ExpressionFormatTest.class, VerificationFormatTest.class, MatchModelTest.class, MatchSimilarityMetricTest.class, SubtitleReaderTestSuite.class, ReleaseInfoTest.class })
 public class FileBotTestSuite {
 	
 }

test/net/sourceforge/filebot/mediainfo/ReleaseInfoTest.java

@@ -0,0 +1,31 @@
+
+package net.sourceforge.filebot.mediainfo;
+
+
+import static org.junit.Assert.*;
+
+import java.io.File;
+
+import org.junit.Test;
+
+
+public class ReleaseInfoTest {
+	
+	@Test
+	public void getVideoSource() {
+		ReleaseInfo info = new ReleaseInfo();
+		File f = new File("Jurassic.Park[1993]DvDrip-aXXo.avi");
+		
+		assertEquals("DvDrip", info.getVideoSource(f));
+	}
+	
+
+	@Test
+	public void getReleaseGroup() throws Exception {
+		ReleaseInfo info = new ReleaseInfo();
+		File f = new File("Jurassic.Park[1993]DvDrip-aXXo.avi");
+		
+		assertEquals("aXXo", info.getReleaseGroup(f));
+	}
+	
+}