fixed Emoji_Modifier_Base is missing 1F46D; cleaned up for easier testing

git-svn-id: https://unicode.org/repos/unicodetools/trunk@1565 13e8329f-0b23-4da4-9fe8-d0f6fe080806

Author: macchiati

unicodetools/org/unicode/tools/emoji/CandidateData.java

@@ -1,9 +1,11 @@
 package org.unicode.tools.emoji;
 
+import java.io.File;
 import java.text.Collator;
 import java.util.ArrayList;
 import java.util.Collections;
 import java.util.Comparator;
+import java.util.Date;
 import java.util.LinkedHashSet;
 import java.util.List;
 import java.util.Locale;
@@ -34,9 +36,11 @@
 import com.ibm.icu.lang.CharSequences;
 import com.ibm.icu.lang.UCharacter;
 import com.ibm.icu.lang.UProperty;
+import com.ibm.icu.text.DateFormat;
 import com.ibm.icu.text.Transform;
 import com.ibm.icu.text.UnicodeSet;
 import com.ibm.icu.util.ICUException;
+import com.ibm.icu.util.ULocale;
 import com.ibm.icu.util.VersionInfo;
 
 public class CandidateData implements Transform<String, String>, EmojiDataSource {
@@ -139,6 +143,7 @@ private CandidateData(String sourceFile) {
         String proposalItem = null;
         Status status = null;
 
+        date = new File(FileUtilities.getRelativeFileName(CandidateData.class, sourceFile)).lastModified();
         for (String line : FileUtilities.in(CandidateData.class, sourceFile)) {
             line = line.trim();
             try {
@@ -491,6 +496,7 @@ public int compare(String o1, String o2) {
             return EmojiOrder.FULL_COMPARATOR.compare(o1, o2);
         }
     };
+    private long date;
 
     /**
      * @return the characters
@@ -944,4 +950,19 @@ public UnicodeSet getAllCharacters(Status status) {
         default: throw new IllegalArgumentException();
         }
     }
+
+    @Override
+    public UnicodeSet getKeycapSequences() {
+        return UnicodeSet.EMPTY;
+    }
+
+    @Override
+    public String addEmojiVariants(String s1) {
+        throw new UnsupportedOperationException();
+    }
+
+    @Override
+    public String getVersionString() {
+        return "candidates:" + DateFormat.getInstanceForSkeleton("yyyyMMdd", ULocale.ROOT).format(date);
+    }
 }

unicodetools/org/unicode/tools/emoji/EmojiData.java

@@ -778,7 +778,7 @@ public UnicodeSet getModifierSequences() {
         return modifierSequences;
     }
 
-    public UnicodeSet getModifiers() {
+    public static UnicodeSet getModifiers() {
         return MODIFIERS;
     }
 
@@ -959,7 +959,7 @@ public boolean skipEmojiSequence(String string) {
 
     private static Pattern EMOJI_VARIANTs = Pattern.compile("[" + Emoji.EMOJI_VARIANT + Emoji.TEXT_VARIANT + "]");
 
-    enum VariantStatus {
+    public enum VariantStatus {
         /** All characters that need them have emoji-variants */
         full("fully-qualified"), 
         /** The first character has an emoji-variant, if needed */
@@ -969,10 +969,18 @@ enum VariantStatus {
         /** Neither full nor partial */
         component("component")
         ;
-        final String name;
+        public final String name;
         private VariantStatus(String name) {
             this.name = name;
         }
+        public static final VariantStatus forString(String name) {
+            for (VariantStatus item : values()) {
+                if (name.equals(item.name)) {
+                    return item;
+                }
+            }
+            return valueOf(name);
+        }
     }
 
     public VariantStatus getVariantStatus(String emoji) {
@@ -1729,6 +1737,10 @@ public UnicodeSet getEmojiWithVariants() {
     public VersionInfo getVersion() {
         return version;
     }
+    
+    public String getVersionString() {
+        return version.getVersionString(2, 2);
+    }
 
     public UnicodeSet getRegionalIndicators() {
         // TODO Auto-generated method stub

unicodetools/org/unicode/tools/emoji/EmojiDataSource.java

@@ -13,6 +13,7 @@ public interface EmojiDataSource {
     public UnicodeSet getExtendedPictographic();
     public UnicodeSet getTagSequences();
     public UnicodeSet getModifierSequences();
+    public UnicodeSet getKeycapSequences();
     public UnicodeSet getFlagSequences();
     public UnicodeSet getZwjSequencesNormal();
     public UnicodeSet getEmojiWithVariants();
@@ -28,6 +29,7 @@ public default String getName(int codepoint) {
         return getName(UTF16.valueOf(codepoint));
     }
     public UnicodeMap<String> getRawNames();
+    
     public default UnicodeSet getBasicSequences() {
         UnicodeSet result = new UnicodeSet();
         for (String s : getSingletonsWithDefectives()) {
@@ -42,5 +44,16 @@ public default UnicodeSet getBasicSequences() {
         }
         return result.freeze();
     }
+    
+    public default UnicodeSet getEmojiForSortRules() {
+        return new UnicodeSet()
+                .addAll(getAllEmojiWithoutDefectives())
+                .removeAll(Emoji.DEFECTIVE)
+                .addAll(getZwjSequencesNormal()) 
+                .addAll(getKeycapSequences());
+    }
+    
+    public String addEmojiVariants(String s1);
+    public String getVersionString();
 }
 

unicodetools/org/unicode/tools/emoji/EmojiDataSourceCombined.java

@@ -120,14 +120,29 @@ public UnicodeMap<String> getRawNames() {
         return add(emojiData.getRawNames(), candidates.getRawNames());
     }
 
-    public static void main(String[] args) {
-        UnicodeSet allChars = EMOJI_DATA.getAllEmojiWithDefectives();
-        
-    }
-
     @Override
     public UnicodeSet getTakesSign() {
         return add(emojiData.getTakesSign(),
                 candidates.getTakesSign());
     }
+
+    @Override
+    public UnicodeSet getKeycapSequences() {
+        return emojiData.getKeycapSequences();
+    }
+
+    @Override
+    public String addEmojiVariants(String s1) {
+        return emojiData.addEmojiVariants(s1);
+    }
+
+    @Override
+    public String getVersionString() {
+        return emojiData.getVersion() + " + " + candidates.getVersionString();
+    }
+    
+//    public static void main(String[] args) {
+//        UnicodeSet allChars = EMOJI_DATA.getAllEmojiWithDefectives();
+//        
+//    }
 }