[ 
https://issues.apache.org/jira/browse/TIKA-179?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12680431#action_12680431
 ] 

Jukka Zitting commented on TIKA-179:
------------------------------------

Also, even the following works (as documented in --help):

    cat PDF.pdf | java -cp target/tika-0.3-standalone.jar 
org.apache.tika.cli.TikaCLI --text -

Let's file an improvement request to make the pipe mode work also without the 
"-".

> Tika stand alone CLI --text output mostly not working, other output formats 
> are fine
> ------------------------------------------------------------------------------------
>
>                 Key: TIKA-179
>                 URL: https://issues.apache.org/jira/browse/TIKA-179
>             Project: Tika
>          Issue Type: Bug
>          Components: cli
>    Affects Versions: 0.2, 0.3
>         Environment: Java 1.5 (also tried Java 1.6). OS used:  Mac OS X, 
> Linux (CentOS)
>            Reporter: Paul Borgermans
>            Assignee: Jukka Zitting
>             Fix For: 0.3
>
>
> When using Tika standalone jar after mvn install in CLI mode, in most of my 
> test documents (pdf, doc, ppt, odt, ), the plain text output option (-t or 
> --text) does not produce any result. When using the other options (xml, html, 
> metadata), the output is correct. Activating debug mode (-v) does not produce 
> additional info either.
> When using the GUI, dragging and dropping does produce the expected results, 
> also in the plain text tab/window
> I rebuilt tika many times in the past 2 months (cleared .m2 directory every 
> time) from svn (latest revision tried:  724002), the CLI --text result is 
> always the same: usually missing output.
> For now, I use the -x output option chained to html2txt as a workaround, but 
> would prefer to use just tika to convert to plain text (which is used for 
> further indexing in Solr).
> Thanks

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to