refactor to use parameterized tests and code cleanup

twitter · Jun 26, 2017 · e93275151f25330db760bc9fb9972e6ecf4b2114 · e932751
1 parent 5b4f3b8
commit e93275151f25330db760bc9fb9972e6ecf4b2114
diff --git a/.gitignore b/.gitignore
@@ -5,3 +5,4 @@ java/.idea
 /.idea
 twitter-text.iml
 .DS_Store
+  
+*.sw?
diff --git a/java/pom.xml b/java/pom.xml
@@ -56,6 +56,7 @@
        <configuration>
          <source>1.5</source>
          <target>1.5</target>
+  
+          <compilerArgument>-Xlint:all</compilerArgument>
        </configuration>
      </plugin>
      <plugin>
@@ -257,14 +258,14 @@
    <dependency>
      <groupId>junit</groupId>
      <artifactId>junit</artifactId>
-  
-      <version>4.11</version>
+  
+      <version>4.12</version>
      <scope>test</scope>
    </dependency>
    <dependency>
      <groupId>org.yaml</groupId>
      <artifactId>snakeyaml</artifactId>
-  
-      <version>1.12</version>
+  
+      <version>1.18</version>
      <scope>test</scope>
    </dependency>

diff --git a/java/src/com/twitter/Autolink.java b/java/src/com/twitter/Autolink.java
@@ -29,12 +29,12 @@
  /** Default attribute for invisible span tag */
  public static final String DEFAULT_INVISIBLE_TAG_ATTRS = "style='position:absolute;left:-9999px;'";
-  
-  public static interface LinkAttributeModifier {
-  
-    public void modify(Entity entity, Map<String, String> attributes);
-  
-  };
+  
+  public interface LinkAttributeModifier {
+  
+    void modify(Entity entity, Map<String, String> attributes);
+  
+  }
-  
-  public static interface LinkTextModifier {
-  
-    public CharSequence modify(Entity entity, CharSequence text);
+  
+  public interface LinkTextModifier {
+  
+    CharSequence modify(Entity entity, CharSequence text);
  }
  protected String urlClass = null;
@@ -74,6 +74,10 @@ private static CharSequence escapeHTML(CharSequence text) {
  }
  public Autolink() {
+  
+    this(true);
+  
+  }
+  
+
+  
+  public Autolink(boolean noFollow) {
    urlClass = null;
    listClass = DEFAULT_LIST_CLASS;
    usernameClass = DEFAULT_USERNAME_CLASS;
@@ -86,6 +90,7 @@ public Autolink() {
    invisibleTagAttrs = DEFAULT_INVISIBLE_TAG_ATTRS;
    extractor.setExtractURLWithoutProtocol(false);
+  
+    this.noFollow = noFollow;
  }
  public String escapeBrackets(String text) {

diff --git a/java/src/com/twitter/Extractor.java b/java/src/com/twitter/Extractor.java
@@ -1,8 +1,12 @@
 package com.twitter;
-  
-import java.util.*;
-  
-import java.util.regex.*;
+  
+import java.util.ArrayList;
+  
+import java.util.Collections;
+  
+import java.util.Comparator;
+  
+import java.util.Iterator;
+  
+import java.util.List;
+  
+import java.util.regex.Matcher;
 /**
 * A class to extract usernames, lists, hashtags and URLs from Tweet text.
@@ -55,14 +59,10 @@ public boolean equals(Object obj) {
      Entity other = (Entity)obj;
-  
-      if (this.type.equals(other.type) &&
+  
+      return this.type.equals(other.type) &&
          this.start == other.start &&
          this.end == other.end &&
-  
-          this.value.equals(other.value)) {
-  
-        return true;
-  
-      } else {
-  
-        return false;
-  
-      }
+  
+          this.value.equals(other.value);
    }
    @Override
@@ -122,7 +122,7 @@ public Extractor() {
  private void removeOverlappingEntities(List<Entity> entities) {
    // sort by index
-  
-    Collections.<Entity>sort(entities, new Comparator<Entity>() {
+  
+    Collections.sort(entities, new Comparator<Entity>() {
      public int compare(Entity e1, Entity e2) {
        return e1.start - e2.start;
      }

diff --git a/java/src/com/twitter/Regex.java b/java/src/com/twitter/Regex.java
@@ -180,7 +180,7 @@
  private static final String URL_VALID_URL_QUERY_ENDING_CHARS = "[a-z0-9_&=#/]";
  private static final String VALID_URL_PATTERN_STRING =
  "(" +                                                            //  $1 total match
-  
-    "(" + URL_VALID_PRECEEDING_CHARS + ")" +                       //  $2 Preceeding chracter
+  
+    "(" + URL_VALID_PRECEEDING_CHARS + ")" +                       //  $2 Preceding character
    "(" +                                                          //  $3 URL
      "(https?://)?" +                                             //  $4 Protocol (optional)
      "(" + URL_VALID_DOMAIN + ")" +                               //  $5 Domain(s)

diff --git a/java/tests/com/twitter/Benchmark.java b/java/tests/com/twitter/Benchmark.java
@@ -1,7 +1,9 @@
 package com.twitter;
-  
-import java.io.File;
 import java.util.List;
+  
+import java.util.Map;
+  
+
+  
+import static org.junit.Assert.assertEquals;
 /**
 * Micro benchmark for discovering hotspots in our autolinker.
@@ -10,10 +12,10 @@
  private static final int AUTO_LINK_TESTS = 10000;
  private static final int ITERATIONS = 10;
+  
+  private static final Autolink linker = new Autolink(false);
  public double testBenchmarkAutolinking() throws Exception {
-  
-    File yamlFile = new File(conformanceDir, "autolink.yml");
-  
-    List testCases = loadConformanceData(yamlFile, "all");
+  
+    List<Map> testCases = loadConformanceData("autolink.yml", "all");
    autolink(testCases);
    long start = System.currentTimeMillis();
    for (int i = 0; i < AUTO_LINK_TESTS; i++) {
@@ -25,9 +27,16 @@ public double testBenchmarkAutolinking() throws Exception {
    return autolinksPerMS;
  }
+  
+  private void autolink(List<Map> testCases) {
+  
+    for (Map testCase : testCases) {
+  
+      assertEquals((String)testCase.get(KEY_DESCRIPTION),
+  
+          testCase.get(KEY_EXPECTED_OUTPUT),
+  
+          linker.autoLink((String) testCase.get(KEY_INPUT)));
+  
+    }
+  
+  }
+  
+
  public static void main(String[] args) throws Exception {
    Benchmark benchmark = new Benchmark();
-  
-    benchmark.setUp();
    double total = 0;
    double best = Double.MAX_VALUE;
    double worst = 0;
@@ -39,7 +48,6 @@ public static void main(String[] args) throws Exception {
    }
    // Drop worst and best
    total -= best + worst;
-  
-    System.out.println("Average: " + (total/(ITERATIONS - 2)));
-  
-    benchmark.tearDown();
+  
+    System.out.println("Average: " + (total / (ITERATIONS - 2)));
  }
 }