[ 
https://issues.apache.org/jira/browse/TIKA-2265?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15868023#comment-15868023
 ] 

Mike Rodent commented on TIKA-2265:
-----------------------------------

I just had a look at the file in question i.e. content.xml when you unzip the 
.docx.

As I understand things it would appear that MS Word may actually be numbering 
footnotes (if they are "per-page") sort of on-the-fly... and of course this 
makes sense: change the pagination, change the footnote numbering.

But the trouble is, with Tika there is no way you can work out where the soft 
page breaks occur.

The .odt file I used when developing my ODF "patch" class (TIKA-2264) turns out 
not to have "per page" numbering... but I just changed this and extracted the 
content.xml.  In this case it appears that the "text:note-citation" value only 
restarts (i.e. to value "1") in the case of a hard break.  So it isn't really 
"per page" numbering in the true sense in LibreOffice Writer.

Anyway, I suspect it may not be possible to do anything about the "anomalous" 
per-page footnote numbering with .docx files...

> Problem with footnotes/endnotes in Tika.parseToString with MS Word (.docx) 
> files
> --------------------------------------------------------------------------------
>
>                 Key: TIKA-2265
>                 URL: https://issues.apache.org/jira/browse/TIKA-2265
>             Project: Tika
>          Issue Type: Bug
>          Components: parser
>    Affects Versions: 1.14
>         Environment: N/A
>            Reporter: Mike Rodent
>            Assignee: Tim Allison
>            Priority: Minor
>              Labels: newbie
>         Attachments: test.docx, test shorter.docx
>
>
> It seems to be the case that a footnote numbered "1" in the real document 
> will be outputted by Tika.parseToString() as "2" in the footnote reference, 
> and "2" in the corresponding footnote body text.... real footnote "2" becomes 
> "3", "3" becomes "4", etc.  Have not yet looked at source code ... I can't 
> imagine it would be difficult to correct this.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Reply via email to