Graphite: Replace names with non-alphanumeric chars with '-'

metrics-graphite/src/main/java/io/dropwizard/metrics/graphite/GraphiteRabbitMQ.java

@@ -9,15 +9,12 @@
 import java.net.Socket;
 import java.nio.charset.Charset;
 import java.util.concurrent.TimeoutException;
-import java.util.regex.Pattern;
 
 /**
  * A rabbit-mq client to a Carbon server.
  */
 public class GraphiteRabbitMQ implements GraphiteSender {
 
-    private static final Pattern WHITESPACE = Pattern.compile("[\\s]+");
-
     private static final Charset UTF_8 = Charset.forName("UTF-8");
 
     private static final Integer DEFAULT_RABBIT_CONNECTION_TIMEOUT_MS = 500;
@@ -167,7 +164,7 @@ public int getFailures() {
     }
 
     public String sanitize(String s) {
-        return WHITESPACE.matcher(s).replaceAll("-");
+        return GraphiteSanitize.sanitize(s, '-');
     }
 
 }

metrics-graphite/src/main/java/io/dropwizard/metrics/graphite/GraphiteSanitize.java

@@ -0,0 +1,71 @@
+package io.dropwizard.metrics.graphite;
+
+class GraphiteSanitize {
+    /** Replaces all characters from a given string that are not ascii and not alphanumeric
+     *  with a dash */
+    static String sanitize(String string, char replacement) {
+        String replaced = replaceFrom(string, replacement);
+
+        // Consolidate multiple dashes into a single one
+        String result = replaced.replace("--", "-");
+        while (!result.equals(replaced)) {
+            replaced = result;
+            result = replaced.replace("--", "-");
+        }
+
+        // Remove any leading or trailing dashes
+        return strip(result, replacement);
+    }
+
+    /** A char matches when it is a letter or digit and it is ASCII, in Guava terminology,
+     *  this would be CharMatcher.ASCII.and(CharMatcher.JAVA_LETTER_OR_DIGIT).negate() */
+    private static boolean matches(char c) {
+        return !(Character.isLetterOrDigit(c) && c <= '\u007f');
+    }
+
+    /** Replace all characters that we're interested in with a replacement character,
+     *  heavily inspired by the same code in Guava's CharMatcher */
+    private static String replaceFrom(String string, char replacement) {
+        int pos = indexIn(string, 0);
+        if (pos == -1) {
+            return string;
+        }
+        char[] chars = string.toCharArray();
+        chars[pos] = replacement;
+        for (int i = pos + 1; i < chars.length; i++) {
+            if (matches(chars[i])) {
+                chars[i] = replacement;
+            }
+        }
+        return new String(chars).trim();
+    }
+
+    /** Finds the first index (or -1) of a character we're interested in */
+    private static int indexIn(String sequence, int start) {
+        int length = sequence.length();
+        for (int i = start; i < length; i++) {
+            if (matches(sequence.charAt(i))) {
+                return i;
+            }
+        }
+        return -1;
+    }
+
+    /** Strips a given character from the beginning and end of a string,
+     *  heavily inspired by Apache's StringUtils.strip
+     */
+    private static String strip(String str, char strip) {
+        int strLen = str.length();
+        int start = 0;
+        int end = strLen - 1;
+        while (start != strLen && str.charAt(start) == strip) {
+            start++;
+        }
+
+        while (end > start && str.charAt(end) == strip) {
+            end--;
+        }
+
+        return start != 0 || end != strLen - 1 ? str.substring(start, end + 1) : str;
+    }
+}

metrics-graphite/src/main/java/io/dropwizard/metrics/graphite/GraphiteUDP.java

@@ -5,15 +5,12 @@
 import java.nio.ByteBuffer;
 import java.nio.channels.DatagramChannel;
 import java.nio.charset.Charset;
-import java.util.regex.Pattern;
 
 /**
  * A client to a Carbon server using unconnected UDP
  */
 public class GraphiteUDP implements GraphiteSender {
 
-    private static final Pattern WHITESPACE = Pattern.compile("[\\s]+");
-
     private static final Charset UTF_8 = Charset.forName("UTF-8");
 
     private final String hostname;
@@ -111,7 +108,7 @@ public void close() throws IOException {
     }
 
     protected String sanitize(String s) {
-        return WHITESPACE.matcher(s).replaceAll("-");
+        return GraphiteSanitize.sanitize(s, '-');
     }
 
 }

metrics-graphite/src/main/java/io/dropwizard/metrics/graphite/PickledGraphite.java

@@ -18,14 +18,12 @@
 import java.nio.charset.Charset;
 import java.util.LinkedList;
 import java.util.List;
-import java.util.regex.Pattern;
 
 /**
  * A client to a Carbon server that sends all metrics after they have been pickled in configurable sized batches
  */
 public class PickledGraphite implements GraphiteSender {
 
-    private static final Pattern WHITESPACE = Pattern.compile("[\\s]+");
     private static final Charset UTF_8 = Charset.forName("UTF-8");
 
     private static final Logger LOGGER = LoggerFactory.getLogger(PickledGraphite.class);
@@ -360,7 +358,7 @@ static class MetricTuple {
     }
 
     protected String sanitize(String s) {
-        return WHITESPACE.matcher(s).replaceAll("-");
+        return GraphiteSanitize.sanitize(s, '-');
     }
 
 }

metrics-graphite/src/test/java/io/dropwizard/metrics/graphite/GraphiteSanitizeTest.java

@@ -0,0 +1,27 @@
+package io.dropwizard.metrics.graphite;
+
+import org.assertj.core.api.SoftAssertions;
+import org.junit.Test;
+
+public class GraphiteSanitizeTest {
+    @Test
+    public void sanitizeGraphiteValues() {
+        SoftAssertions softly = new SoftAssertions();
+
+        softly.assertThat(GraphiteSanitize.sanitize("Foo Bar", '-')).isEqualTo("Foo-Bar");
+        softly.assertThat(GraphiteSanitize.sanitize(" Foo Bar ", '-')).isEqualTo("Foo-Bar");
+        softly.assertThat(GraphiteSanitize.sanitize(" Foo Bar", '-')).isEqualTo("Foo-Bar");
+        softly.assertThat(GraphiteSanitize.sanitize("Foo Bar ", '-')).isEqualTo("Foo-Bar");
+        softly.assertThat(GraphiteSanitize.sanitize("  Foo Bar  ", '-')).isEqualTo("Foo-Bar");
+        softly.assertThat(GraphiteSanitize.sanitize("Foo@Bar", '-')).isEqualTo("Foo-Bar");
+        softly.assertThat(GraphiteSanitize.sanitize("Foó Bar", '-')).isEqualTo("Fo-Bar");
+        softly.assertThat(GraphiteSanitize.sanitize("||ó/.", '-')).isEqualTo("");
+        softly.assertThat(GraphiteSanitize.sanitize("${Foo:Bar:baz}", '-')).isEqualTo("Foo-Bar-baz");
+        softly.assertThat(GraphiteSanitize.sanitize("St. Foo's of Bar", '-')).isEqualTo("St-Foo-s-of-Bar");
+        softly.assertThat(GraphiteSanitize.sanitize("(Foo and (Bar and (Baz)))", '-')).isEqualTo("Foo-and-Bar-and-Baz");
+        softly.assertThat(GraphiteSanitize.sanitize("Foo.bar.baz", '-')).isEqualTo("Foo-bar-baz");
+        softly.assertThat(GraphiteSanitize.sanitize("FooBar", '-')).isEqualTo("FooBar");
+
+        softly.assertAll();
+    }
+}