Add nextStringMethod

neo4j · May 11, 2018 · aa59147 · aa59147
1 parent bc02024
commit aa59147
Show file tree

Hide file tree

Showing 3 changed files with 59 additions and 16 deletions.
diff --git a/community/random-values/src/main/java/org/neo4j/values/RandomValue.java b/community/random-values/src/main/java/org/neo4j/values/RandomValue.java
@@ -193,6 +193,38 @@ public TextValue nextAlphaNumericString( int minLength, int maxLength )
         return Values.utf8Value( bytes );
     }
 
+    public TextValue nextString( int minLength, int maxLength )
+    {
+        int length = intBetween( minLength, maxLength );
+        UTF8StringValueBuilder builder = new UTF8StringValueBuilder( nextPowerOf2( length ) );
+
+        for ( int i = 0; i < length; i++ )
+        {
+            boolean validCodePoint = false;
+            while ( !validCodePoint )
+            {
+                int codePoint = intBetween( Character.MIN_CODE_POINT, Character.MAX_CODE_POINT );
+                switch ( Character.getType( codePoint ) )
+                {
+                case Character.UNASSIGNED:
+                case Character.PRIVATE_USE:
+                case Character.SURROGATE:
+                    continue;
+                default:
+                    builder.addCodePoint( codePoint );
+                    validCodePoint = true;
+                }
+            }
+        }
+        return builder.build();
+    }
+
+    private int nextPowerOf2( int i )
+    {
+        return 1 << (32 - Integer.numberOfLeadingZeros( i ));
+    }
+
+
     private int intBetween( int min, int max )
     {
         return min + random.nextInt( max - min + 1 );

diff --git a/community/random-values/src/main/java/org/neo4j/values/UTF8StringValueBuilder.java b/community/random-values/src/main/java/org/neo4j/values/UTF8StringValueBuilder.java
@@ -64,40 +64,40 @@ TextValue build()
         return Values.utf8Value( bytes, 0, length );
     }
 
-    void addCodePoint( int codepoint )
+    void addCodePoint( int codePoint )
     {
-        assert codepoint >= 0;
+        assert codePoint >= 0;
-        if ( codepoint < 0x80 )
+        if ( codePoint < 0x80 )
         {
             //one byte is all it takes
-            add( (byte) codepoint );
+            add( (byte) codePoint );
         }
-        else if ( codepoint < 0x800 )
+        else if ( codePoint < 0x800 )
         {
             //Require two bytes - will be laid out like:
             //b1       b2
             //110xxxxx 10xxxxxx
-            add( (byte) (0b1100_0000 | (0b0001_1111 & (codepoint >> 6))) );
+            add( (byte) (0b1100_0000 | (0b0001_1111 & (codePoint >> 6))) );
-            add( (byte) (0b1000_0000 | (0b0011_1111 & codepoint)) );
+            add( (byte) (0b1000_0000 | (0b0011_1111 & codePoint)) );
         }
-        else if ( codepoint < 0x10000 )
+        else if ( codePoint < 0x10000 )
         {
             //Require three bytes - will be laid out like:
             //b1       b2       b3
             //1110xxxx 10xxxxxx 10xxxxxx
-            add( (byte) (0b1110_0000 | (0b0000_1111 & (codepoint >> 12))) );
+            add( (byte) (0b1110_0000 | (0b0000_1111 & (codePoint >> 12))) );
-            add( (byte) (0b1000_0000 | (0b0011_1111 & (codepoint >> 6))) );
+            add( (byte) (0b1000_0000 | (0b0011_1111 & (codePoint >> 6))) );
-            add( (byte) (0b1000_0000 | (0b0011_1111 & codepoint)) );
+            add( (byte) (0b1000_0000 | (0b0011_1111 & codePoint)) );
         }
         else
         {
             //Require four bytes - will be laid out like:
             //b1       b2       b3
             //11110xxx 10xxxxxx 10xxxxxx
-            add( (byte) (0b1111_0000 | (0b0001_1111 & (codepoint >> 18))) );
+            add( (byte) (0b1111_0000 | (0b0001_1111 & (codePoint >> 18))) );
-            add( (byte) (0b1000_0000 | (0b0011_1111 & (codepoint >> 12))) );
+            add( (byte) (0b1000_0000 | (0b0011_1111 & (codePoint >> 12))) );
-            add( (byte) (0b1000_0000 | (0b0011_1111 & (codepoint >> 6))) );
+            add( (byte) (0b1000_0000 | (0b0011_1111 & (codePoint >> 6))) );
-            add( (byte) (0b1000_0000 | (0b0011_1111 & codepoint)) );
+            add( (byte) (0b1000_0000 | (0b0011_1111 & codePoint)) );
         }
     }
 }
diff --git a/community/random-values/src/test/java/org/neo4j/values/RandomValueTest.java b/community/random-values/src/test/java/org/neo4j/values/RandomValueTest.java
@@ -120,7 +120,6 @@ public void nextShortValueBounded()
         checkBounded( () -> randomValue.nextShortValue( BOUND ) );
     }
 
-
     @Test
     public void nextByteValueUnbounded()
     {
@@ -216,6 +215,18 @@ public void nextAlphaNumericString()
         assertThat( seenDigits, empty() );
     }
 
+    @Test
+    public void nextString()
+    {
+        for ( int i = 0; i < ITERATIONS; i++ )
+        {
+            TextValue textValue = randomValue.nextString( 10, 20 );
+            String asString = textValue.stringValue();
+            int length = asString.codePointCount( 0, asString.length() );
+            assertThat( length, greaterThanOrEqualTo( 10 ) );
+            assertThat( length, lessThanOrEqualTo( 20 ) );
+        }
+    }
 
     private void checkDistribution( Supplier<Value> supplier )
     {