package net.sourceforge.filebot.media; import static java.util.Arrays.*; import static java.util.ResourceBundle.*; import static java.util.regex.Pattern.*; import static net.sourceforge.filebot.similarity.Normalization.*; import static net.sourceforge.tuned.FileUtilities.*; import static net.sourceforge.tuned.StringUtilities.*; import java.io.File; import java.io.FileFilter; import java.io.IOException; import java.io.InputStreamReader; import java.nio.ByteBuffer; import java.nio.charset.Charset; import java.text.Collator; import java.text.Normalizer; import java.text.Normalizer.Form; import java.util.ArrayList; import java.util.Collection; import java.util.Comparator; import java.util.HashSet; import java.util.List; import java.util.Locale; import java.util.Map; import java.util.Scanner; import java.util.Set; import java.util.TreeMap; import java.util.regex.Matcher; import java.util.regex.Pattern; import java.util.zip.GZIPInputStream; import net.sourceforge.filebot.web.CachedResource; import net.sourceforge.filebot.web.Movie; import net.sourceforge.tuned.ByteBufferInputStream; public class ReleaseInfo { public String getVideoSource(File file) { // check parent and itself for group names return matchLast(getVideoSourcePattern(), getBundle(getClass().getName()).getString("pattern.video.source").split("[|]"), file.getParent(), file.getName()); } public String getReleaseGroup(File file) throws IOException { // check parent and itself for group names return matchLast(getReleaseGroupPattern(false), releaseGroupResource.get(), file.getParent(), file.getName()); } public Locale getLanguageSuffix(String name) { // match locale identifier and lookup Locale object Map languages = getLanguageMap(Locale.ENGLISH, Locale.getDefault()); String lang = matchLast(getLanguageSuffixPattern(languages.keySet()), null, name); if (lang == null) return null; return languages.get(lang); } protected String matchLast(Pattern pattern, String[] standardValues, CharSequence... sequence) { String lastMatch = null; // match last occurrence for (CharSequence name : sequence) { if (name == null) continue; Matcher matcher = pattern.matcher(name); while (matcher.find()) { lastMatch = matcher.group(); } } // prefer standard value over matched value if (lastMatch != null && standardValues != null) { for (String standard : standardValues) { if (standard.equalsIgnoreCase(lastMatch)) { return standard; } } } return lastMatch; } public List cleanRelease(Collection items, boolean strict) throws IOException { Set languages = getLanguageMap(Locale.ENGLISH, Locale.getDefault()).keySet(); Pattern releaseGroup = getReleaseGroupPattern(strict); Pattern languageSuffix = getLanguageSuffixPattern(languages); Pattern languageTag = getLanguageTagPattern(languages); Pattern videoSource = getVideoSourcePattern(); Pattern videoFormat = getVideoFormatPattern(); Pattern resolution = getResolutionPattern(); Pattern queryBlacklist = getBlacklistPattern(); Pattern[] blacklist = new Pattern[] { releaseGroup, languageSuffix, languageTag, videoSource, videoFormat, resolution, queryBlacklist }; Pattern[] stopwords = new Pattern[] { getReleaseGroupPattern(true), languageSuffix, languageTag, videoSource, videoFormat, resolution }; List output = new ArrayList(items.size()); for (String it : items) { it = substringBefore(it, stopwords); it = clean(it, blacklist); // ignore empty values if (it.length() > 0) { output.add(it); } } return output; } public String clean(String item, Pattern... blacklisted) { for (Pattern it : blacklisted) { item = it.matcher(item).replaceAll(""); } return normalizePunctuation(item); } public String substringBefore(String item, Pattern... stopwords) { for (Pattern it : stopwords) { Matcher matcher = it.matcher(item); if (matcher.find()) { return item.substring(0, matcher.start()); // use substring before the matched stopword } } // no stopword found, keep original string return item; } public Pattern getLanguageTagPattern(Collection languages) { // [en] return compile("(?<=[-\\[{(])(" + join(quoteAll(languages), "|") + ")(?=\\p{Punct})", CASE_INSENSITIVE | UNICODE_CASE | CANON_EQ); } public Pattern getLanguageSuffixPattern(Collection languages) { // .en.srt return compile("(?<=\\p{Punct}|\\s)(" + join(quoteAll(languages), "|") + ")(?=$)", CASE_INSENSITIVE | UNICODE_CASE | CANON_EQ); } public Pattern getResolutionPattern() { // match screen resolutions 640x480, 1280x720, etc return compile("(? releaseGroupResource = new PatternResource(getBundle(getClass().getName()).getString("url.release-groups")); protected final CachedResource queryBlacklistResource = new PatternResource(getBundle(getClass().getName()).getString("url.query-blacklist")); protected final CachedResource movieListResource = new MovieResource(getBundle(getClass().getName()).getString("url.movie-list")); protected final CachedResource seriesListResource = new SeriesResource(getBundle(getClass().getName()).getString("url.series-list")); protected static class PatternResource extends CachedResource { public PatternResource(String resource) { super(resource, String[].class, 24 * 60 * 60 * 1000); // 24h update interval } @Override public String[] process(ByteBuffer data) { return compile("\\n").split(Charset.forName("UTF-8").decode(data)); } } protected static class MovieResource extends CachedResource { public MovieResource(String resource) { super(resource, Movie[].class, 7 * 24 * 60 * 60 * 1000); // check for updates once a week } @Override public Movie[] process(ByteBuffer data) throws IOException { Scanner scanner = new Scanner(new GZIPInputStream(new ByteBufferInputStream(data)), "UTF-8").useDelimiter("\t|\n"); List movies = new ArrayList(); while (scanner.hasNext()) { int imdbid = scanner.nextInt(); String name = scanner.next(); int year = scanner.nextInt(); movies.add(new Movie(name, year, imdbid)); } return movies.toArray(new Movie[0]); } } protected static class SeriesResource extends CachedResource { public SeriesResource(String resource) { super(resource, String[].class, 7 * 24 * 60 * 60 * 1000); // check for updates once a week } @Override public String[] process(ByteBuffer data) throws IOException { return readAll(new InputStreamReader(new GZIPInputStream(new ByteBufferInputStream(data)), "utf-8")).split("\\n"); } } protected static class FolderEntryFilter implements FileFilter { private final Pattern entryPattern; public FolderEntryFilter(Pattern entryPattern) { this.entryPattern = entryPattern; } @Override public boolean accept(File dir) { if (dir.isDirectory()) { for (String entry : dir.list()) { if (entryPattern.matcher(entry).matches()) { return true; } } } return false; } } private Collection quoteAll(Collection strings) { List patterns = new ArrayList(strings.size()); for (String it : strings) { patterns.add(Pattern.quote(it)); } return patterns; } private Map getLanguageMap(Locale... supportedDisplayLocale) { // use maximum strength collator by default Collator collator = Collator.getInstance(Locale.ROOT); collator.setDecomposition(Collator.FULL_DECOMPOSITION); collator.setStrength(Collator.PRIMARY); @SuppressWarnings("unchecked") Comparator order = (Comparator) collator; Map languageMap = new TreeMap(order); Set displayLocales = new HashSet(asList(supportedDisplayLocale)); for (String code : Locale.getISOLanguages()) { Locale locale = new Locale(code); languageMap.put(locale.getLanguage(), locale); languageMap.put(locale.getISO3Language(), locale); // map display language names for given locales for (Locale language : displayLocales) { // make sure language name is properly normalized so accents and whatever don't break the regex pattern syntax String languageName = Normalizer.normalize(locale.getDisplayLanguage(language), Form.NFKD); languageMap.put(languageName, locale); } } // remove illegal tokens languageMap.remove(""); return languageMap; } }