TIKA-4749 - improve inline handling of metadata only (#2866)

Author: tballison

tika-core/src/main/java/org/apache/tika/parser/AutoDetectParser.java

@@ -160,6 +160,10 @@ public void parse(TikaInputStream tis, ContentHandler handler, Metadata metadata
         // don't leak into CONTENT_TYPE
         metadata.set(Metadata.CONTENT_TYPE,
                 EmbeddedDocumentUtil.normalizeMediaType(type.toString()));
+        // Metadata-only pseudo-parse: register the entry, skip the content parse.
+        if (context.get(MetadataOnlyParse.class) != null) {
+            return;
+        }
         //check for zero-byte inputstream
         if (tis.getOpenContainer() == null) {
             if (autoDetectParserConfig.getThrowOnZeroBytes()) {

tika-core/src/main/java/org/apache/tika/parser/MetadataOnlyParse.java

@@ -0,0 +1,35 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You 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 org.apache.tika.parser;
+
+/**
+ * ParseContext marker telling {@link AutoDetectParser} to register the embedded
+ * entry but skip the content parse. Set by metadata-only passes that pseudo-parse
+ * a placeholder stream only to register an entry. Independent of throwOnZeroBytes.
+ */
+public final class MetadataOnlyParse {
+
+    /**
+     * Singleton instance indicating the current parse should not dispatch to a
+     * content parser.
+     */
+    public static final MetadataOnlyParse INSTANCE = new MetadataOnlyParse();
+
+    private MetadataOnlyParse() {
+        // Private constructor for singleton
+    }
+}

tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-pdf-module/src/main/java/org/apache/tika/parser/pdf/image/ImageGraphicsEngine.java

@@ -57,14 +57,14 @@
 
 import org.apache.tika.exception.TikaException;
 import org.apache.tika.exception.TikaMemoryLimitException;
-import org.apache.tika.exception.ZeroByteFileException;
 import org.apache.tika.extractor.EmbeddedDocumentExtractor;
 import org.apache.tika.extractor.EmbeddedDocumentUtil;
 import org.apache.tika.io.BoundedInputStream;
 import org.apache.tika.io.TikaInputStream;
 import org.apache.tika.metadata.Metadata;
 import org.apache.tika.metadata.TikaCoreProperties;
 import org.apache.tika.metadata.TikaPagedText;
+import org.apache.tika.parser.MetadataOnlyParse;
 import org.apache.tika.parser.ParseContext;
 import org.apache.tika.parser.pdf.PDFParserConfig;
 import org.apache.tika.parser.pdf.PDMetadataExtractor;
@@ -448,16 +448,14 @@ protected void extractInlineImageMetadataOnly(PDImage pdImage, Metadata metadata
         metadata.set(Metadata.IMAGE_WIDTH, pdImage.getWidth());
         metadata.set(Metadata.IMAGE_LENGTH, pdImage.getHeight());
         //TODO: what else can we extract from the PDImage without rendering?
-        ZeroByteFileException.IgnoreZeroByteFileException before =
-                parseContext.get(ZeroByteFileException.IgnoreZeroByteFileException.class);
+        //Register the image's metadata entry without decoding it (marker skips the parse).
         try (TikaInputStream tis = TikaInputStream.get(new byte[0])) {
-            parseContext.set(ZeroByteFileException.IgnoreZeroByteFileException.class,
-                    ZeroByteFileException.IGNORE_ZERO_BYTE_FILE_EXCEPTION);
+            parseContext.set(MetadataOnlyParse.class, MetadataOnlyParse.INSTANCE);
             embeddedDocumentExtractor.parseEmbedded(tis,
                     new EmbeddedContentHandler(xhtml), metadata, parseContext, false);
         } finally {
-            //replace whatever was there before
-            parseContext.set(ZeroByteFileException.IgnoreZeroByteFileException.class, before);
+            //clear so it can't leak to the next image
+            parseContext.set(MetadataOnlyParse.class, null);
         }
     }
 

tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-pdf-module/src/test/java/org/apache/tika/parser/pdf/PDFParserTest.java

@@ -47,7 +47,6 @@
 import org.apache.tika.config.loader.TikaLoader;
 import org.apache.tika.exception.AccessPermissionException;
 import org.apache.tika.exception.EncryptedDocumentException;
-import org.apache.tika.exception.ZeroByteFileException;
 import org.apache.tika.extractor.DocumentSelector;
 import org.apache.tika.io.TikaInputStream;
 import org.apache.tika.metadata.Font;
@@ -62,7 +61,9 @@
 import org.apache.tika.metadata.XMPPDF;
 import org.apache.tika.mime.MediaType;
 import org.apache.tika.parser.AutoDetectParser;
+import org.apache.tika.parser.AutoDetectParserConfig;
 import org.apache.tika.parser.CompositeParser;
+import org.apache.tika.parser.MetadataOnlyParse;
 import org.apache.tika.parser.ParseContext;
 import org.apache.tika.parser.Parser;
 import org.apache.tika.parser.PasswordProvider;
@@ -1358,7 +1359,7 @@ public void testExtractInlineImageMetadata() throws Exception {
         config.setExtractInlineImageMetadataOnly(true);
         context.set(PDFParserConfig.class, config);
         List<Metadata> metadataList = getRecursiveMetadata("testOCR.pdf", context);
-        assertNull(context.get(ZeroByteFileException.IgnoreZeroByteFileException.class));
+        assertNull(context.get(MetadataOnlyParse.class));
         assertEquals(2, metadataList.size());
         assertEquals("image/png", metadataList.get(1).get(Metadata.CONTENT_TYPE));
         assertEquals("/image-0.png",
@@ -1368,6 +1369,37 @@ public void testExtractInlineImageMetadata() throws Exception {
         assertEquals("image-0.png", metadataList.get(1).get(TikaCoreProperties.RESOURCE_NAME_KEY));
     }
 
+    @Test
+    public void testExtractInlineImageMetadataThrowOnZeroBytesFalse() throws Exception {
+        //TIKA-4749: in metadata-only mode the inline image is registered via a
+        //placeholder pseudo-parse. With throwOnZeroBytes=false that placeholder used
+        //to be handed to a real parser (image/OCR), recording a spurious embedded
+        //exception. The MetadataOnlyParse marker must make it skip the parse instead.
+        ParseContext context = new ParseContext();
+        PDFParserConfig config = new PDFParserConfig();
+        config.setExtractInlineImageMetadataOnly(true);
+        context.set(PDFParserConfig.class, config);
+
+        AutoDetectParser p = new AutoDetectParser();
+        AutoDetectParserConfig adpc = new AutoDetectParserConfig();
+        adpc.setThrowOnZeroBytes(false);
+        p.setAutoDetectParserConfig(adpc);
+
+        List<Metadata> metadataList =
+                getRecursiveMetadata("testOCR.pdf", p, new Metadata(), context, false);
+        assertNull(context.get(MetadataOnlyParse.class));
+        assertEquals(2, metadataList.size());
+        Metadata image = metadataList.get(1);
+        assertEquals("image/png", image.get(Metadata.CONTENT_TYPE));
+        assertEquals(261, (int) image.getInt(Metadata.IMAGE_LENGTH));
+        assertEquals(934, (int) image.getInt(Metadata.IMAGE_WIDTH));
+        //the placeholder must not be dispatched to any content parser. Without the
+        //fix it is (EmptyParser here; ImageParser+TesseractOCRParser when tesseract
+        //is installed, which is what records the spurious embedded exception).
+        assertEquals(0, image.getValues(TikaCoreProperties.TIKA_PARSED_BY).length);
+        assertNull(image.get(TikaCoreProperties.EMBEDDED_EXCEPTION));
+    }
+
     /**
      * Simple class to count end of document events.  If functionality is useful,
      * move to org.apache.tika in src/test