From cb10a44ce4558ce1ac8eed10ce8d9719104021dc Mon Sep 17 00:00:00 2001
From: Reinhard Pointner <rednoah@users.noreply.github.com>
Date: Wed, 19 Feb 2014 20:28:00 +0000
Subject: [PATCH] * improve {group} name extraction

---
 .../filebot/format/MediaBindingBean.java      | 20 ++++++++++++++++++-
 .../filebot/media/ReleaseInfo.java            |  4 ++++
 .../filebot/similarity/Normalization.java     |  4 ++++
 website/data/release-groups.txt               |  1 +
 4 files changed, 28 insertions(+), 1 deletion(-)

diff --git a/source/net/sourceforge/filebot/format/MediaBindingBean.java b/source/net/sourceforge/filebot/format/MediaBindingBean.java
index 0752b5f4..69375c32 100644
--- a/source/net/sourceforge/filebot/format/MediaBindingBean.java
+++ b/source/net/sourceforge/filebot/format/MediaBindingBean.java
@@ -419,8 +419,26 @@ public class MediaBindingBean {
 		// use inferred media file
 		File inferredMediaFile = getInferredMediaFile();
 
+		// consider foldername, filename and original filename
+		String[] filenames = new String[] { inferredMediaFile.getParentFile().getName(), inferredMediaFile.getName(), getOriginalFileName(inferredMediaFile) };
+
+		// reduce false positives by removing the know titles from the name
+		List<String> titles = new ArrayList<String>();
+		titles.add(getName());
+		titles.add(getYear().toString());
+		titles.addAll(getAliasNames());
+		Pattern nonGroupPattern = releaseInfo.getCustomRemovePattern(titles);
+
+		for (int i = 0; i < filenames.length; i++) {
+			if (filenames[i] == null)
+				continue;
+
+			// normalize space characters
+			filenames[i] = nonGroupPattern.matcher(normalizeSpace(filenames[i], " ")).replaceAll("");
+		}
+
 		// look for release group names in media file and it's parent folder
-		return releaseInfo.getReleaseGroup(inferredMediaFile.getParent(), inferredMediaFile.getName(), getOriginalFileName(inferredMediaFile));
+		return releaseInfo.getReleaseGroup(filenames);
 	}
 
 	@Define("lang")
diff --git a/source/net/sourceforge/filebot/media/ReleaseInfo.java b/source/net/sourceforge/filebot/media/ReleaseInfo.java
index 7a31ff68..f343c8e8 100644
--- a/source/net/sourceforge/filebot/media/ReleaseInfo.java
+++ b/source/net/sourceforge/filebot/media/ReleaseInfo.java
@@ -266,6 +266,10 @@ public class ReleaseInfo {
 		return compile(join(excludeBlacklistResource.get(), "|"), CASE_INSENSITIVE | UNICODE_CASE);
 	}
 
+	public Pattern getCustomRemovePattern(Collection<String> terms) throws IOException {
+		return compile("(?<!\\p{Alnum})(" + join(quoteAll(terms), "|") + ")(?!\\p{Alnum})", CASE_INSENSITIVE | UNICODE_CASE);
+	}
+
 	public Movie[] getMovieList() throws IOException {
 		return movieListResource.get();
 	}
diff --git a/source/net/sourceforge/filebot/similarity/Normalization.java b/source/net/sourceforge/filebot/similarity/Normalization.java
index 0e52a3d4..c83d21b5 100644
--- a/source/net/sourceforge/filebot/similarity/Normalization.java
+++ b/source/net/sourceforge/filebot/similarity/Normalization.java
@@ -30,6 +30,10 @@ public class Normalization {
 		return name;
 	}
 
+	public static String normalizeSpace(String name, String replacement) {
+		return name.replaceAll("[:?._]", " ").trim().replaceAll("\\s+", replacement);
+	}
+
 	public static String removeEmbeddedChecksum(String string) {
 		// match embedded checksum and surrounding brackets
 		return checksum.matcher(string).replaceAll("");
diff --git a/website/data/release-groups.txt b/website/data/release-groups.txt
index da4fe638..95213545 100644
--- a/website/data/release-groups.txt
+++ b/website/data/release-groups.txt
@@ -2003,6 +2003,7 @@ SATANiC
 SATIVA
 SAVANNAH
 SB-Subs
+SbR
 SCARED
 SChiZO
 Scratch404