From 411a7bd508f5d90d5adb988d238b9c99f7a3d66a Mon Sep 17 00:00:00 2001 From: "Cezar.Leahu" Date: Fri, 13 Sep 2019 12:50:37 +0300 Subject: [PATCH] ATS-515: Default options for TIKA when called through ATS - add default "UTF-8" target encoding - restore previously default values for the other options --- .../org/alfresco/transformer/TikaController.java | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/alfresco-docker-tika/src/main/java/org/alfresco/transformer/TikaController.java b/alfresco-docker-tika/src/main/java/org/alfresco/transformer/TikaController.java index d223f0c0..4b0182a3 100644 --- a/alfresco-docker-tika/src/main/java/org/alfresco/transformer/TikaController.java +++ b/alfresco-docker-tika/src/main/java/org/alfresco/transformer/TikaController.java @@ -26,6 +26,7 @@ */ package org.alfresco.transformer; +import static java.lang.Boolean.parseBoolean; import static org.alfresco.transformer.executors.Tika.INCLUDE_CONTENTS; import static org.alfresco.transformer.executors.Tika.NOT_EXTRACT_BOOKMARKS_TEXT; import static org.alfresco.transformer.executors.Tika.PDF_BOX; @@ -36,7 +37,6 @@ import static org.alfresco.transformer.fs.FileManager.createSourceFile; import static org.alfresco.transformer.fs.FileManager.createTargetFile; import static org.alfresco.transformer.fs.FileManager.createTargetFileName; import static org.alfresco.transformer.util.MimetypeMap.MIMETYPE_TEXT_PLAIN; -import static org.alfresco.transformer.util.Util.stringToBoolean; import static org.springframework.http.HttpStatus.OK; import static org.springframework.http.MediaType.MULTIPART_FORM_DATA_VALUE; @@ -173,16 +173,18 @@ public class TikaController extends AbstractTransformerController logger.debug("Processing request with: sourceFile '{}', targetFile '{}', transformOptions" + " '{}', timeout {} ms", sourceFile, targetFile, transformOptions, timeout); - final Boolean includeContents = stringToBoolean(transformOptions.get("includeContents")); - final Boolean notExtractBookmarksText = stringToBoolean( - transformOptions.get("notExtractBookmarksText")); - final String targetEncoding = transformOptions.get("targetEncoding"); + final boolean includeContents = parseBoolean( + transformOptions.getOrDefault("includeContents", "false")); + final boolean notExtractBookmarksText = parseBoolean( + transformOptions.getOrDefault("notExtractBookmarksText", "false")); + final String targetEncoding = transformOptions.getOrDefault("targetEncoding", "UTF-8"); final String transform = getTransformerName(sourceFile, sourceMimetype, targetMimetype, transformOptions); + javaExecutor.call(sourceFile, targetFile, transform, - includeContents != null && includeContents ? INCLUDE_CONTENTS : null, - notExtractBookmarksText != null && notExtractBookmarksText ? NOT_EXTRACT_BOOKMARKS_TEXT : null, + includeContents ? INCLUDE_CONTENTS : null, + notExtractBookmarksText ? NOT_EXTRACT_BOOKMARKS_TEXT : null, TARGET_MIMETYPE + targetMimetype, TARGET_ENCODING + targetEncoding); } }