You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tika.apache.org by "Finn Woelm (JIRA)" <ji...@apache.org> on 2019/01/18 06:52:00 UTC

[jira] [Commented] (TIKA-1975) Different behaviour between tika-app and tika-server

    [ https://issues.apache.org/jira/browse/TIKA-1975?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16745838#comment-16745838 ] 

Finn Woelm commented on TIKA-1975:
----------------------------------

Issue also exists in version 1.20

This also affects things like images in the DOCX file (see sample file below):
Server response: "[image: ]Hello world"
Client response: "Hello world"

Is there a way to change the server behavior to match the client (eg by making some changes and building from source)?

[^word-with-image.docx]

> Different behaviour between tika-app and tika-server
> ----------------------------------------------------
>
>                 Key: TIKA-1975
>                 URL: https://issues.apache.org/jira/browse/TIKA-1975
>             Project: Tika
>          Issue Type: Bug
>          Components: general
>    Affects Versions: 1.12
>            Reporter: Sebastian Hesse
>            Priority: Major
>         Attachments: word-with-image.docx
>
>
> I am using tika-server to extract content from word files via DotNet.
> For the extraction i use the following rest endpoint (https://wiki.apache.org/tika/TikaJAXRS#Get_the_Text_of_a_Document).
> If I extract the content of a DOCX file the content contains some hidden bookmarks like: "[bookmark:_GoBack] hello world"
> When i do the same with the tika-app via console i get "hello world"
> I didn't find a way to prevent tika-server from extracting the hidden bookmarks. Also specifying the mime-type did not work.
> Here is a test file (only a few chars) http://en.file-upload.net/download-11584028/ContentWord.docx.html



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)