[
https://issues.apache.org/jira/browse/TIKA-3006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16992276#comment-16992276
]
Tilman Hausherr commented on TIKA-3006:
---------------------------------------
I ran tika app and here's what I got:
Current (1.24 rc2):
Keywords: keyword1, keyword2
meta:keyword: keyword1, keyword2
pdf:docinfo:keywords: keyword1, keyword2
1.22:
Keywords: keyword1, keyword2
dc:subject: keyword1, keyword2
meta:keyword: keyword1, keyword2
pdf:docinfo:keywords: keyword1, keyword2
Technically the current output is correct, assuming that "dc" means "dublin
core". This would be in a PDF with a metadata stream, and your PDF does not
have that. I had a quick look at the code... Until TIKA-2917
extractMultilingualItems() added data from the /Info dictionary when metadata
wasn't available.
> Regression in PDF keywords extraction since 1.23
> ------------------------------------------------
>
> Key: TIKA-3006
> URL: https://issues.apache.org/jira/browse/TIKA-3006
> Project: Tika
> Issue Type: Bug
> Reporter: David Pilato
> Priority: Major
> Attachments: test.pdf
>
>
> Hey team.
>
> I have a unit test which is not passing anymore with Tika 1.23. Code is
> [https://github.com/dadoonet/fscrawler/blob/master/tika/src/test/java/fr/pilato/elasticsearch/crawler/fs/tika/TikaDocParserTest.java#L430-L496].
>
> Basically, this line is now failing:
> {{assertThat(doc.getMeta().getKeywords(), containsInAnyOrder("keyword1", "
> keyword2"));}}
>
> But also this line which seems related:
> {{assertThat(raw, hasEntry("dc:subject", "keyword1, keyword2"));}}
>
> This line is failing as well although not related I think:
> {{assertThat(raw, hasEntry("created", "2016-07-07T08:37:42Z"));}}
>
> I'm not sure if it's a bug or expected. If the later, I'll simply change my
> test suite.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
