Merge pull request #675 from dadoonet/fix/674-warn-error-tika

Warn in case of Tika error
dadoonet · Feb 1, 2019 · feea255 · feea255
2 parents c48f908 + 95b37e0
commit feea255
Show file tree

Hide file tree

Showing 2 changed files with 21 additions and 1 deletion.
diff --git a/tika/src/main/java/fr/pilato/elasticsearch/crawler/fs/tika/TikaDocParser.java b/tika/src/main/java/fr/pilato/elasticsearch/crawler/fs/tika/TikaDocParser.java
@@ -93,6 +93,16 @@ public static void generate(FsSettings fsSettings, InputStream inputStream, Stri
                 parsedContent = extractText(fsSettings, indexedChars, inputStream, metadata);
                 logger.trace("End of Tika extraction");
             } catch (Throwable e) {
+                // Build a message from embedded errors
+                Throwable current = e;
+                StringBuilder sb = new StringBuilder();
+                while (current != null) {
+                    sb.append(" -> ");
+                    sb.append(current.getMessage());
+                    current = current.getCause();
+                }
+
+                logger.warn("Failed to extract [" + indexedChars + "] characters of text for [" + filename + "] {}", sb.toString());
                 logger.debug("Failed to extract [" + indexedChars + "] characters of text for [" + filename + "]", e);
             }
 

diff --git a/tika/src/test/java/fr/pilato/elasticsearch/crawler/fs/tika/TikaDocParserTest.java b/tika/src/test/java/fr/pilato/elasticsearch/crawler/fs/tika/TikaDocParserTest.java
@@ -697,8 +697,17 @@ public void testShiftJisEncoding() throws IOException {
         assertThat(doc.getContent(), not(isEmptyOrNullString()));
     }
 
+    /**
+     * Test protected document
+     */
+    @Test
+    public void testProtectedDocument() throws IOException {
+        FsSettings fsSettings = FsSettings.builder(getCurrentTestName()).build();
+        Doc doc = extractFromFile("test-protected.docx", fsSettings);
+        assertThat(doc.getFile().getContentType(), is("application/x-tika-ooxml-protected"));
+    }
+
     private Doc extractFromFileExtension(String extension) throws IOException {
-        logger.info("Test extraction of [{}] file", extension);
         return extractFromFile("test." + extension);
     }
 
@@ -707,6 +716,7 @@ private Doc extractFromFile(String filename) throws IOException {
     }
 
     private Doc extractFromFile(String filename, FsSettings fsSettings) throws IOException {
+        logger.info("Test extraction of [{}]", filename);
         InputStream data = getBinaryContent(filename);
         Doc doc = new Doc();
         MessageDigest messageDigest = null;