2.3.1

change the tika fork parser that takes a filename input and modify it slightly so that instead of sending the
socket the content input stream, send the filename so that the tika fork main can use the FileInputStream directly.
This will cut out some unnecessary socket communication.

CHANGELOG.txt

@@ -1,3 +1,9 @@
+2.3.1
+
+change the tika fork parser that takes a filename input and modify it slightly so that instead of sending the
+socket the content input stream, send the filename so that the tika fork main can use the FileInputStream directly.
+This will cut out some unnecessary socket communication.
+
 2.3
 
 bump to tika 1.24.1 instead of using the custom build

gradle.properties

@@ -1,5 +1,5 @@
 # version
-version=2.3
+version=2.3.1
 
 # depsre
 tikaVersion=1.24.1

tika-fork-client/src/main/java/org/apache/tika/client/TikaRunner.java

@@ -5,8 +5,8 @@
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
+import java.io.ByteArrayInputStream;
 import java.io.EOFException;
-import java.io.FileInputStream;
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.ObjectInputStream;
@@ -54,14 +54,14 @@ public Metadata parse(String baseUri,
                         OutputStream contentOutputStream,
                         long abortAfterMs,
                         long maxBytesToParse) throws InterruptedException, ExecutionException, TimeoutException, IOException {
-    try (FileInputStream fis = new FileInputStream(filename)) {
-      return parse(baseUri,
-          contentType,
-          fis,
-          contentOutputStream,
-          abortAfterMs,
-          maxBytesToParse);
-    }
+    ByteArrayInputStream bais = new ByteArrayInputStream(filename.getBytes());
+    return parseImpl(baseUri,
+        contentType,
+        bais,
+        contentOutputStream,
+        abortAfterMs,
+        maxBytesToParse,
+        true);
   }
 
   public Metadata parse(String baseUri,
@@ -70,11 +70,27 @@ public Metadata parse(String baseUri,
                         OutputStream contentOutputStream,
                         long abortAfterMs,
                         long maxBytesToParse) throws InterruptedException, ExecutionException, TimeoutException {
+    return parseImpl(baseUri,
+        contentType,
+        contentInStream,
+        contentOutputStream,
+        abortAfterMs,
+        maxBytesToParse,
+        false);
+  }
+
+  private Metadata parseImpl(String baseUri,
+                             String contentType,
+                             InputStream contentInStream,
+                             OutputStream contentOutputStream,
+                             long abortAfterMs,
+                             long maxBytesToParse,
+                             boolean inputIsFilename) throws InterruptedException, ExecutionException, TimeoutException {
     ExecutorService es = Executors.newFixedThreadPool(3, new TikaRunnerThreadFactory());
     try {
       es.submit(() -> {
         try {
-          writeContent(baseUri, contentType, contentInPort, contentInStream);
+          writeContent(baseUri, contentType, contentInPort, contentInStream, inputIsFilename);
         } catch (Exception e) {
           throw new RuntimeException("Failed to send content stream to forked Tika parser JVM", e);
         }
@@ -140,13 +156,16 @@ public Metadata parse(String baseUri,
   private void writeContent(String baseUri,
                             String contentType,
                             int port,
-                            InputStream contentInStream) throws Exception {
+                            InputStream contentInStream,
+                            boolean inputIsFilename) throws Exception {
     Socket socket = getSocket(InetAddress.getLocalHost().getHostAddress(), port);
     try (OutputStream out = socket.getOutputStream()) {
       out.write(baseUri.getBytes());
       out.write('\n');
       out.write(contentType.getBytes());
       out.write('\n');
+      out.write(String.valueOf(inputIsFilename).getBytes());
+      out.write('\n');
       long numChars;
       do {
         numChars = IOUtils.copy(contentInStream, out);

tika-fork-main/src/main/java/org/apache/tika/fork/main/TikaForkMain.java

@@ -21,8 +21,10 @@
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 import org.xml.sax.ContentHandler;
+import org.xml.sax.SAXException;
 
 import java.io.File;
+import java.io.FileInputStream;
 import java.io.FileReader;
 import java.io.IOException;
 import java.io.InputStream;
@@ -325,31 +327,46 @@ private void parseFile(OutputStream metadataOutputStream, OutputStream contentOu
 
       String baseUri = "";
       String contentType = "";
+      String inputIsFilenameStr = "";
       int nextChar;
       while ((nextChar = inputStream.read()) != '\n') {
         baseUri += (char)nextChar;
       }
       while ((nextChar = inputStream.read()) != '\n') {
         contentType += (char)nextChar;
       }
-
-      LOG.info("Next file to parse baseUri={}, contentType={}", baseUri, contentType);
+      while ((nextChar = inputStream.read()) != '\n') {
+        inputIsFilenameStr += (char)nextChar;
+      }
 
       if (StringUtils.isNotBlank(contentType)) {
         metadata.set(Metadata.CONTENT_TYPE, contentType);
       }
 
-      TikaInputStream tikaInputStream = TikaInputStream.get(inputStream);
+      if (Boolean.parseBoolean(inputIsFilenameStr)) {
+        String tmpFilePath = IOUtils.toString(inputStream);
+        try (FileInputStream fis = new FileInputStream(tmpFilePath)) {
+          LOG.info("Next file to parse baseUri={}, contentType={}, tmpFilePath={}", baseUri, contentType, tmpFilePath);
+          doTikaParse(contentOutputStream, context, metadata, compositeParser, objectOutputStream, baseUri, TikaInputStream.get(fis));
+        }
+      } else {
+        LOG.info("Next file to parse baseUri={}, contentType={}", baseUri, contentType);
 
-      TikaParsingHandler contentHandler = getContentHandler(baseUri, contentOutputStream, extractHtmlLinks);
-      compositeParser.parse(tikaInputStream, contentHandler, metadata, context);
+        doTikaParse(contentOutputStream, context, metadata, compositeParser, objectOutputStream, baseUri, TikaInputStream.get(inputStream));
+      }
 
-      objectOutputStream.writeObject(metadata);
     } finally {
       contentOutputStream.close();
     }
   }
 
+  private void doTikaParse(OutputStream contentOutputStream, ParseContext context, Metadata metadata, CompositeParser compositeParser, ObjectOutputStream objectOutputStream, String baseUri, TikaInputStream tikaInputStream2) throws TikaException, IOException, SAXException {
+    TikaInputStream tikaInputStream = tikaInputStream2;
+    TikaParsingHandler contentHandler = getContentHandler(baseUri, contentOutputStream, extractHtmlLinks);
+    compositeParser.parse(tikaInputStream, contentHandler, metadata, context);
+    objectOutputStream.writeObject(metadata);
+  }
+
   /**
    * Runs the external tika parsing server.
    */