Andrzej Bialecki created TIKA-1413:
---------------------------------------
Summary: OOXML thumbnail name added to body
Key: TIKA-1413
URL: https://issues.apache.org/jira/browse/TIKA-1413
Project: Tika
Issue Type: Bug
Components: parser
Affects Versions: 1.6
Reporter: Andrzej Bialecki
AbstractOOXMLExtractor.handleThumbnail processes thumbnails using
EmbeddedDocumentExtractor, but with the outputHtml flag set to true (unlike
other embedded parts in handleEmbeddedParts(...)).
This results in adding the thumbnail name to the main body of the document (as
a package-entry), which in my opinion is wrong.
Example:
{code}
<?xml version="1.0" encoding="UTF-8"?><html
xmlns="http://www.w3.org/1999/xhtml";>
<head>
<meta name="meta:slide-count" content="1"/>
<meta name="cp:revision" content="5"/>
<meta name="meta:last-author" content="Nick Burch"/>
<meta name="Slide-Count" content="1"/>
<meta name="Last-Author" content="Nick Burch"/>
<meta name="meta:save-date" content="2010-09-08T16:15:14Z"/>
<meta name="Content-Length" content="202969"/>
<meta name="subject" content="Gym class featuring a brown fox and lazy dog"/>
<meta name="Application-Name" content="Microsoft Office PowerPoint"/>
<meta name="Author" content="Nevin Nollop"/>
<meta name="dcterms:created" content="1601-01-01T00:00:00Z"/>
<meta name="Application-Version" content="12.0000"/>
<meta name="date" content="2010-09-08T16:15:14Z"/>
<meta name="Total-Time" content="2"/>
<meta name="extended-properties:Template" content=""/>
<meta name="publisher" content=""/>
<meta name="creator" content="Nevin Nollop"/>
<meta name="Word-Count" content="9"/>
<meta name="meta:paragraph-count" content="1"/>
<meta name="extended-properties:AppVersion" content="12.0000"/>
<meta name="Creation-Date" content="1601-01-01T00:00:00Z"/>
<meta name="meta:author" content="Nevin Nollop"/>
<meta name="cp:subject" content="Gym class featuring a brown fox and lazy dog"/>
<meta name="extended-properties:Application" content="Microsoft Office
PowerPoint"/>
<meta name="resourceName" content="testPPT_embeded.pptx"/>
<meta name="Paragraph-Count" content="1"/>
<meta name="dc:title" content="The quick brown fox jumps over the lazy dog"/>
<meta name="Last-Save-Date" content="2010-09-08T16:15:14Z"/>
<meta name="custom:Version" content="1"/>
<meta name="Revision-Number" content="5"/>
<meta name="Last-Printed" content="1601-01-01T00:00:00Z"/>
<meta name="meta:print-date" content="1601-01-01T00:00:00Z"/>
<meta name="meta:creation-date" content="1601-01-01T00:00:00Z"/>
<meta name="dcterms:modified" content="2010-09-08T16:15:14Z"/>
<meta name="Template" content=""/>
<meta name="dc:creator" content="Nevin Nollop"/>
<meta name="meta:word-count" content="9"/>
<meta name="extended-properties:Company" content=""/>
<meta name="Last-Modified" content="2010-09-08T16:15:14Z"/>
<meta name="extended-properties:PresentationFormat" content="On-screen Show
(4:3)"/>
<meta name="X-Parsed-By" content="org.apache.tika.parser.DefaultParser"/>
<meta name="X-Parsed-By"
content="org.apache.tika.parser.microsoft.ooxml.OOXMLParser"/>
<meta name="modified" content="2010-09-08T16:15:14Z"/>
<meta name="xmpTPg:NPages" content="1"/>
<meta name="extended-properties:TotalTime" content="2"/>
<meta name="dc:publisher" content=""/>
<meta name="Content-Type"
content="application/vnd.openxmlformats-officedocument.presentationml.presentation"/>
<meta name="Presentation-Format" content="On-screen Show (4:3)"/>
<title>The quick brown fox jumps over the lazy dog</title>
</head>
<body><p>The quick brown fox jumps over the lazy dog</p>
<div class="embedded" id="slide1_rId4"/>
<div class="embedded" id="slide1_rId5"/>
<div class="embedded" id="slide1_rId6"/>
<div class="embedded" id="slide1_rId7"/>
<div class="embedded" id="slide1_rId8"/>
<div class="embedded" id="slide1_rId9"/>
<div class="embedded" id="thumbnail_0.jpeg"/><div
class="package-entry"><h1>thumbnail_0.jpeg</h1></div></body></html>
{code}
The extracted plain text looks like this (using tika-app):
{code}
The quick brown fox jumps over the lazy dog
thumbnail_0.jpeg
{code}
The fix is trivial - change the flag in AbstractOOXMLExtractor:158 to false.
I think also that the id attribute should be set to the real thumbnail path
within the package (i.e. tPart.getPartName().getName()) instead of the
artificially created sequential name.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
