[jira] [Reopened] (TIKA-2510) Embedded MP3 file in PPTX document no longer identified

["Tim Allison (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/TIKA-2510?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Tim Allison reopened TIKA-2510:
-------------------------------

Figured out diff btwn 1.14 and 1.16. We added the set {{seen}} so that we only processed a media item once.  In 1.16, 

1. the first relationship to the mp3 file was type {{http://schemas.microsoft.com/office/2007/relationships/media}}
2. we added the mp3 to {{seen}}
3. we then didn't process it because we didn't recognize that relationship
4. the second relationship to the mp3 file was type {{http://schemas.openxmlformats.org/officeDocument/2006/relationships/audio}}
5. In 1.14, we processed this, but in 1.16, because we had added that media item to seen (without processing it!), we skipped it.

The correct fix is to add embedded media items to {{seen}} only after they are processed.

> Embedded MP3 file in PPTX document no longer identified
> -------------------------------------------------------
>
>                 Key: TIKA-2510
>                 URL: https://issues.apache.org/jira/browse/TIKA-2510
>             Project: Tika
>          Issue Type: Bug
>          Components: parser
>    Affects Versions: 1.15
>            Reporter: Eamonn Saunders
>            Assignee: Tim Allison
>            Priority: Minor
>             Fix For: 1.17
>
>         Attachments: Windows Audio File.pptx, tika-1.14-output.json, tika-1.15-output.json
>
>
> I'm attaching a sample PPTX file with an embedded MP3 file along with JSON files produced by Tika App (versions 1.14 and 1.15).
> Notice that the 1.14 output identifies the embedded MP3 file while the 1.15 version does not.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)