Add test case for IndexHnswDenseVectors (#2254)

+ Add custom log appender to intercept and verify logging output
+ Clean up IndexInvertedDenseVectors
+ Fix code coverage (wasn't running properly before)
+ Other light clean-up

.github/workflows/maven.yml

@@ -26,4 +26,5 @@ jobs:
       run: mvn -B package --file pom.xml
     - name: Upload coverage to Codecov
       uses: codecov/codecov-action@v3
-      # https://github.com/codecov/codecov-action
+      with:
+        token: ${{ secrets.CODECOV_TOKEN }}

pom.xml

@@ -93,7 +93,8 @@
         <artifactId>maven-surefire-plugin</artifactId>
         <version>3.1.0</version>
         <configuration>
-          <argLine>-Xmx8192m</argLine>
+          <!-- Need to add @{argLine} for JaCoCo to work; see https://stackoverflow.com/questions/71568474/getting-skipping-jacoco-execution-due-to-missing-execution-data-file-with-jaco -->
+          <argLine>@{argLine} -Xmx8192m</argLine>
         </configuration>
       </plugin>
       <plugin>

src/main/java/io/anserini/eval/RelevanceJudgments.java

@@ -34,7 +34,7 @@
 public class RelevanceJudgments {
   final private Map<String, Map<String, Integer>> qrels;
   final private static String CACHE_DIR = Paths.get(System.getProperty("user.home"), "/.cache/anserini/topics-and-qrels").toString();
-  final private static String CLOUD_PATH = "https://raw.githubusercontent.com/castorini/anserini-tools/master/topics-and-qrels/";
+  final private static String SERVER_PATH = "https://raw.githubusercontent.com/castorini/anserini-tools/master/topics-and-qrels/";
 
   public static RelevanceJudgments fromQrels(Qrels qrels) throws IOException {
     return new RelevanceJudgments("src/main/resources/" + qrels.path);
@@ -46,7 +46,7 @@ public RelevanceJudgments(String file) throws IOException {
     try {
       qrelsPath = getQrelsPath(qrelsPath);
     } catch (IOException e) {
-      System.out.println("Qrels file not found at " + qrelsPath.toString());
+      System.out.println("Qrels file not found at " + qrelsPath);
     }
 
     try (BufferedReader br = new BufferedReader(new FileReader(qrelsPath.toString()))) {
@@ -66,7 +66,7 @@ public RelevanceJudgments(String file) throws IOException {
         }
       }
     } catch (IOException e) {
-      throw new IOException("Could not read qrels file");
+      throw new IOException("Could not read qrels file!");
     }
   }
 
@@ -136,7 +136,7 @@ public static String getQrelsResource(Path qrelsPath) throws IOException {
     try {
       resultPath = getQrelsPath(qrelsPath);
     } catch (Exception e) {
-      throw new IOException("Could not get qrels file from cloud or local file system");
+      throw new IOException("Could not get qrels file either from server or local file system!");
     }
 
     InputStream inputStream = Files.newInputStream(resultPath);
@@ -184,13 +184,13 @@ public static Path getNewQrelAbsPath(Path qrelsPath) {
    * @throws IOException if qrels file is not found
    */
   public static Path downloadQrels(Path qrelsPath) throws IOException {
-    String qrelsURL = CLOUD_PATH + qrelsPath.getFileName().toString().toString();
-    System.out.println("Downloading qrels from cloud " + qrelsURL.toString());
+    String qrelsURL = SERVER_PATH + qrelsPath.getFileName().toString();
+    System.out.println("Downloading qrels from " + qrelsURL);
     File qrelsFile = new File(getCacheDir(), qrelsPath.getFileName().toString());
     try {
       FileUtils.copyURLToFile(new URL(qrelsURL), qrelsFile);
     } catch (Exception e) {
-      System.out.println("Error downloading topics from cloud " + qrelsURL.toString());
+      System.out.println("Error downloading topics from " + qrelsURL);
       throw e;
     }
     return qrelsFile.toPath();

src/main/java/io/anserini/index/IndexInvertedDenseVectors.java

@@ -546,15 +546,13 @@ public Counters run() throws IOException {
       }
 
       writer.commit();
-      System.out.println(String.format("%s docs indexed", counters.indexed.get()));
-      long space = FileUtils.sizeOfDirectory(indexPath.toFile()) / (1024L * 1024L);
-      System.out.println(String.format("Index size: %dMB", space));
       writer.close();
 
-      final long durationMillis =
-          TimeUnit.MILLISECONDS.convert(System.nanoTime() - start, TimeUnit.NANOSECONDS);
-      System.out.println(String.format("Total time: %s",
-                                       DurationFormatUtils.formatDuration(durationMillis, "HH:mm:ss")));
+      LOG.info(String.format("Indexing Complete! %,d documents indexed", counters.indexed.get()));
+
+      final long durationMillis = TimeUnit.MILLISECONDS.convert(System.nanoTime() - start, TimeUnit.NANOSECONDS);
+      LOG.info(String.format("Total %,d documents indexed in %s", counters.indexed.get(),
+          DurationFormatUtils.formatDuration(durationMillis, "HH:mm:ss")));
     }
 
     return counters;

src/main/java/io/anserini/index/SimpleIndexer.java

@@ -280,46 +280,4 @@ public void close(boolean optimize) {
       }
     }
   }
-
-  // Main here exists only as a convenience. Once we're happy with the APIs, there should *not* be
-  // a separate code entry point; one less thing to debug, one less thing to go wrong.
-  public static void main(String[] argv) throws Exception {
-    Args args = new Args();
-    CmdLineParser parser = new CmdLineParser(args, ParserProperties.defaults().withUsageWidth(100));
-
-    try {
-      parser.parseArgument(argv);
-    } catch (CmdLineException e) {
-      System.err.println(e.getMessage());
-      parser.printUsage(System.err);
-      System.err.println("Example: " + SimpleIndexer.class.getSimpleName() +
-          parser.printExample(OptionHandlerFilter.REQUIRED));
-      return;
-    }
-
-    final long start = System.nanoTime();
-    JsonCollection collection = new JsonCollection(Paths.get(args.input));
-
-    int cnt = 0;
-    SimpleIndexer indexer = new SimpleIndexer(args);
-
-    LOG.info("input: " + args.input);
-    LOG.info("collection: " + args.index);
-
-    for (FileSegment<JsonCollection.Document> segment : collection ) {
-      for (JsonCollection.Document doc : segment) {
-        indexer.addRawDocument(doc.raw());
-        cnt++;
-        if (cnt % 100000 == 0) {
-          LOG.info(cnt + " docs indexed");
-        }
-      }
-      segment.close();
-    }
-
-    indexer.close();
-    final long durationMillis = TimeUnit.MILLISECONDS.convert(System.nanoTime() - start, TimeUnit.NANOSECONDS);
-    LOG.info(String.format("Total %,d documents indexed in %s", cnt,
-        DurationFormatUtils.formatDuration(durationMillis, "HH:mm:ss")));
-  }
 }

src/main/java/io/anserini/search/topicreader/TopicReader.java

@@ -45,7 +45,7 @@ public abstract class TopicReader<K> {
   protected final int BUFFER_SIZE = 1 << 16; // 64K
   protected Path topicFile;
   final private static String CACHE_DIR = Paths.get(System.getProperty("user.home"), "/.cache/anserini/topics-and-qrels").toString();
-  final private static String CLOUD_PATH = "https://raw.githubusercontent.com/castorini/anserini-tools/master/topics-and-qrels/";
+  final private static String SERVER_PATH = "https://raw.githubusercontent.com/castorini/anserini-tools/master/topics-and-qrels/";
 
 
   static private final Map<String, Class<? extends TopicReader>> TOPIC_FILE_TO_TYPE = new HashMap<>();
@@ -232,13 +232,13 @@ private static String getCacheDir() {
    * @throws IOException if error encountered downloading topics
    */
   public static Path getTopicsFromCloud(Path topicPath) throws IOException{
-    String topicURL = CLOUD_PATH + topicPath.getFileName().toString();
-    System.out.println("Downloading topics from cloud " + topicURL.toString());
+    String topicURL = SERVER_PATH + topicPath.getFileName().toString();
+    System.out.println("Downloading topics from " + topicURL);
     File topicFile = new File(getCacheDir(), topicPath.getFileName().toString());
     try{
       FileUtils.copyURLToFile(new URL(topicURL), topicFile);
     }catch (Exception e){
-      System.out.println("Error downloading topics from cloud " + topicURL.toString());
+      System.out.println("Error downloading topics from " + topicURL);
       throw e;
     }
     return topicFile.toPath();

src/test/java/io/anserini/CustomAppender.java

@@ -0,0 +1,37 @@
+/*
+ * Anserini: A Lucene toolkit for reproducible information retrieval research
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package io.anserini;
+
+import org.apache.logging.log4j.core.LogEvent;
+import org.apache.logging.log4j.core.appender.AbstractAppender;
+
+public class CustomAppender extends AbstractAppender {
+  private String lastLog = null;
+
+  public CustomAppender(String name) {
+    super(name, null, null, true, null);
+  }
+
+  public String getLastLog() {
+    return lastLog;
+  }
+
+  @Override
+  public void append(LogEvent event) {
+    lastLog = event.getMessage().getFormattedMessage();
+  }
+}

src/test/java/io/anserini/collection/Iso19115CollectionTest.java

@@ -43,7 +43,6 @@ public void setUp() throws Exception {
 
   @Override
   void checkDocument(SourceDocument doc, Map<String, String> expected) {
-    // System.out.println(((Iso19115Collection.Document) doc).getThesaurusName());
     assertTrue(doc.indexable());
     assertEquals(expected.get("id"), doc.id());
     assertEquals(expected.get("title"), ((Iso19115Collection.Document) doc).getTitle());

src/test/java/io/anserini/index/IndexHnswDenseVectorsTest.java

@@ -0,0 +1,78 @@
+/*
+ * Anserini: A Lucene toolkit for reproducible information retrieval research
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package io.anserini.index;
+
+import io.anserini.CustomAppender;
+import org.apache.logging.log4j.Level;
+import org.apache.logging.log4j.LogManager;
+import org.apache.logging.log4j.Logger;
+import org.apache.logging.log4j.core.config.Configurator;
+import org.junit.AfterClass;
+import org.junit.BeforeClass;
+import org.junit.Test;
+
+import java.util.LinkedList;
+import java.util.List;
+
+import static org.junit.Assert.assertTrue;
+
+/**
+ * Tests for {@link IndexHnswDenseVectors}
+ */
+public class IndexHnswDenseVectorsTest {
+  private static final Logger LOGGER = LogManager.getLogger(IndexHnswDenseVectors.class);
+  private static CustomAppender APPENDER;
+
+  @BeforeClass
+  public static void setupClass() {
+    APPENDER = new CustomAppender("CustomAppender");
+    APPENDER.start();
+
+    ((org.apache.logging.log4j.core.Logger) LOGGER).addAppender(APPENDER);
+
+    Configurator.setLevel(IndexHnswDenseVectors.class.getName(), Level.INFO);
+  }
+
+  @Test
+  public void test1() throws Exception {
+    List<String> args = new LinkedList<>();
+    args.add("-collection");
+    args.add("JsonDenseVectorCollection");
+    args.add("-input");
+    args.add("src/test/resources/sample_docs/openai_ada2/json_vector");
+    args.add("-index");
+    args.add("target/idx-sample-hnsw" + System.currentTimeMillis());
+    args.add("-generator");
+    args.add("LuceneDenseVectorDocumentGenerator");
+    args.add("-threads");
+    args.add("1");
+    args.add("-M");
+    args.add("16");
+    args.add("-efC");
+    args.add("100");
+
+    IndexHnswDenseVectors.main(args.toArray(new String[0]));
+
+    System.out.println(APPENDER.getLastLog());
+    assertTrue(APPENDER.getLastLog().contains("Total 100 documents indexed"));
+  }
+
+  @AfterClass
+  public static void teardownClass() {
+    ((org.apache.logging.log4j.core.Logger) LOGGER).removeAppender(APPENDER);
+  }
+}