[
https://issues.apache.org/jira/browse/TIKA-4891?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18114786#comment-18114786
]
Tilman Hausherr edited comment on TIKA-4891 at 9/13/26 12:23 PM:
-----------------------------------------------------------------
I'm not experienced at reviewing AI code, and clicking on AI to reviewing AI
code sounds funny (I might still do it later 😂).
So here's my spontaneous thoughts how we could evaluate this:
- deleted TODOs
- my TODO ("should be merged with COSDictionary segment below") and the weird
findPages() method has been deleted 👍
- however I can't find a check for "confirm the right number of pages was
found"
but I assume this is indirectly done through the new AUTO / TAGS settings
- deleted / modified tests. The new code should succeed at all old tests and
more.
I notice some of the tests are modified and the closing tags are missing, this
should be investigated.
The good thing is that I didn't see deleted tests.
- do a human compare on some interesting PDFs
- Test with a PDF that has ActualText, there are some test files in
PDFBOX-5868 and in related issues.
I'll do a local build and see what this does.
was (Author: tilman):
I'm not experienced at reviewing AI code, and clicking on AI to reviewing AI
code sounds funny (I might still do it later 😂).
So here's my spontaneous thoughts how we could evaluate this:
- deleted TODOs
- my TODO ("should be merged with COSDictionary segment below") and the weird
findPages() method has been deleted 👍
- however I can't find a check for "confirm the right number of pages was
found"
but I assume this is indirectly done through the new AUTO / TAGS settings
- deleted / modified tests. The new code should succeed at all old tests and
more.
I notice some of the tests are modified and the closing tags are missing,
this should be investigated.
The good thing is that I didn't see deleted tests.
- do a human compare on some interesting PDFs
- Test with a PDF that has ActualText, there are some test files in PDFBOX-5868
and in related issues.
> Improve handling of PDF/UA structural tags/marked content
> ---------------------------------------------------------
>
> Key: TIKA-4891
> URL: https://issues.apache.org/jira/browse/TIKA-4891
> Project: Tika
> Issue Type: Task
> Reporter: Tim Allison
> Priority: Major
> Attachments: 1008690.pdf, image-2026-09-13-10-54-59-251.png,
> screenshot-1.png
>
>
> PDF/UA includes structural markup. We hacked out a standalone handler for
> this back in 1.x but haven't touched it in years.
>
> We should modernize our handling of structural tags and eventually consider
> turning that on by default. That decision will be based on evaluation on
> 1000s of PDFs. This is not a default switch to be taken lightly.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
