Make categories into an enum. Show only printable characters.

2 files changed, 90 insertions, 7 deletions

diff --git a/src/org/madore/android/unicodeMap/UnicodeCharacter.java b/src/org/madore/android/unicodeMap/UnicodeCharacter.java
index fb68169..11d8d3e 100644
--- a/src/org/madore/android/unicodeMap/UnicodeCharacter.java
+++ b/src/org/madore/android/unicodeMap/UnicodeCharacter.java
@@ -1,14 +1,89 @@
 package org.madore.android.unicodeMap;
 
+import java.util.Arrays;
+import java.util.Set;
+import java.util.Map;
+import java.util.EnumSet;
+import java.util.HashMap;
 import java.util.Formatter;
 
 public class UnicodeCharacter {
 
+    enum Category {
+	UPPERCASE_LETTER("Lu", Character.UPPERCASE_LETTER),
+	LOWERCASE_LETTER("Ll", Character.LOWERCASE_LETTER),
+	TITLECASE_LETTER("Lt", Character.TITLECASE_LETTER),
+	MODIFIER_LETTER("Lm", Character.MODIFIER_LETTER),
+	OTHER_LETTER("Lo", Character.OTHER_LETTER),
+	NON_SPACING_MARK("Mn", Character.NON_SPACING_MARK),
+	COMBINING_SPACING_MARK("Mc", Character.COMBINING_SPACING_MARK),
+	ENCLOSING_MARK("Me", Character.ENCLOSING_MARK),
+	DECIMAL_DIGIT_NUMBER("Nd", Character.DECIMAL_DIGIT_NUMBER),
+	LETTER_NUMBER("Nl", Character.LETTER_NUMBER),
+	OTHER_NUMBER("No", Character.OTHER_NUMBER),
+	CONNECTOR_PUNCTUATION("Pc", Character.CONNECTOR_PUNCTUATION),
+	DASH_PUNCTUATION("Pd", Character.DASH_PUNCTUATION),
+	START_PUNCTUATION("Ps", Character.START_PUNCTUATION),
+	END_PUNCTUATION("Pe", Character.END_PUNCTUATION),
+	INITIAL_QUOTE_PUNCTUATION("Pi", Character.INITIAL_QUOTE_PUNCTUATION),
+	FINAL_QUOTE_PUNCTUATION("Pf", Character.FINAL_QUOTE_PUNCTUATION),
+	OTHER_PUNCTUATION("Po", Character.OTHER_PUNCTUATION),
+	MATH_SYMBOL("Sm", Character.MATH_SYMBOL),
+	CURRENCY_SYMBOL("Sc", Character.CURRENCY_SYMBOL),
+	MODIFIER_SYMBOL("Sk", Character.MODIFIER_SYMBOL),
+	OTHER_SYMBOL("So", Character.OTHER_SYMBOL),
+	SPACE_SEPARATOR("Zs", Character.SPACE_SEPARATOR),
+	LINE_SEPARATOR("Zl", Character.LINE_SEPARATOR),
+	PARAGRAPH_SEPARATOR("Zp", Character.PARAGRAPH_SEPARATOR),
+	CONTROL("Cc", Character.CONTROL),
+	FORMAT("Cf", Character.FORMAT),
+	SURROGATE("Cs", Character.SURROGATE),
+	PRIVATE_USE("Co", Character.PRIVATE_USE),
+	UNASSIGNED("Cn", Character.UNASSIGNED);
+	public final String code;
+	public final byte javaValue;
+	Category(String code, byte javaValue) {
+	    this.code = code;
+	    this.javaValue = javaValue;
+	}
+	protected final static Map<String,Category> revMap
+	    = new HashMap<String,Category>();
+	static {
+	    for ( Category cat : Category.values() )
+		revMap.put(cat.code, cat);
+	}
+	static Category fromCode(String code) {
+	    Category cat = revMap.get(code);
+	    if ( cat == null )
+		cat = UNASSIGNED;
+	    return cat;
+	}
+    }
+
+    protected final static Set<Category> printable;
+
+    static {
+	Category[] prlist = new Category[] {
+	    Category.UPPERCASE_LETTER, Category.LOWERCASE_LETTER,
+	    Category.TITLECASE_LETTER, Category.MODIFIER_LETTER,
+	    Category.OTHER_LETTER, Category.NON_SPACING_MARK,
+	    Category.COMBINING_SPACING_MARK, Category.ENCLOSING_MARK,
+	    Category.DECIMAL_DIGIT_NUMBER, Category.LETTER_NUMBER,
+	    Category.OTHER_NUMBER, Category.CONNECTOR_PUNCTUATION,
+	    Category.DASH_PUNCTUATION, Category.START_PUNCTUATION,
+	    Category.END_PUNCTUATION, Category.INITIAL_QUOTE_PUNCTUATION,
+	    Category.FINAL_QUOTE_PUNCTUATION, Category.OTHER_PUNCTUATION,
+	    Category.MATH_SYMBOL, Category.CURRENCY_SYMBOL,
+	    Category.MODIFIER_SYMBOL, Category.OTHER_SYMBOL
+	};
+	printable = EnumSet.copyOf(Arrays.asList(prlist));
+    }
+
     protected final int codePoint;
     protected final String name;
-    protected final String category;
+    protected final Category category;
 
-    public UnicodeCharacter(int codePoint, String name, String category) {
+    public UnicodeCharacter(int codePoint, String name, Category category) {
 	this.codePoint = codePoint;
 	this.name = name;
 	this.category = category;
@@ -22,15 +97,22 @@ public class UnicodeCharacter {
 	return name;
     }
 
+    public boolean isPrintable() {
+	return printable.contains(this.category);
+    }
+
     @Override
     public String toString() {
 	StringBuilder s = new StringBuilder();
-	s.append(Character.toChars(codePoint));
+	if ( this.isPrintable() ) {
+	    s.append(Character.toChars(codePoint));
+	    s.append(": ");
+	}
 	Formatter fmt = new Formatter(s);
 	if ( codePoint < 0x10000 )
-	    fmt.format(" (U+%04X %s)", codePoint, name);
+	    fmt.format("U+%04X %s", codePoint, name);
 	else
-	    fmt.format(" (U+%X %s)", codePoint, name);
+	    fmt.format("U+%X %s", codePoint, name);
 	return new String(s);
     }
 
diff --git a/src/org/madore/android/unicodeMap/UnicodeDatabase.java b/src/org/madore/android/unicodeMap/UnicodeDatabase.java
index 9e809f2..9dff87f 100644
--- a/src/org/madore/android/unicodeMap/UnicodeDatabase.java
+++ b/src/org/madore/android/unicodeMap/UnicodeDatabase.java
@@ -101,7 +101,8 @@ public class UnicodeDatabase {
 	if ( c.getCount() > 0 ) {
 	    c.moveToFirst();
 	    ch = new UnicodeCharacter(c.getInt(0),
-				      c.getString(1), c.getString(2));
+				      c.getString(1),
+				      UnicodeCharacter.Category.fromCode(c.getString(2)));
 	} else
 	    ch = null;
 	c.close();
@@ -144,7 +145,7 @@ public class UnicodeDatabase {
 		    if ( hasNext ) {
 			UnicodeCharacter ch
 			    = new UnicodeCharacter(c.getInt(0), c.getString(1),
-						   c.getString(2));
+						   UnicodeCharacter.Category.fromCode(c.getString(2)));
 			if ( c.isLast() ) {
 			    hasNext = false;
 			    c.close();