jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jukka Zitting (JIRA)" <j...@apache.org>
Subject [jira] Commented: (JCR-1878) Use Apache Tika for text extraction
Date Mon, 24 Aug 2009 10:22:59 GMT

    [ https://issues.apache.org/jira/browse/JCR-1878?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12746804#action_12746804

Jukka Zitting commented on JCR-1878:

I upgraded the Tika dependency to 0.4 in revision 807139. This dropped the size of jackrabbit-standalone
jar from 44MB to 39MB. It's still quite a bit, but until we have an alternative way to extract
text from OOXML files we're stuck with the large ooxml-schemas jar. I had a discussion with
the POI team about a trimmed down "extract only" version of the OOXML support, but it looks
like that would have to use a different design than the current implementation and there doesn't
seem to be too many people with such an itch. So for now I'm OK to live with the extra megabytes.

> Use Apache Tika for text extraction
> -----------------------------------
>                 Key: JCR-1878
>                 URL: https://issues.apache.org/jira/browse/JCR-1878
>             Project: Jackrabbit Content Repository
>          Issue Type: Improvement
>          Components: jackrabbit-text-extractors
>            Reporter: Jukka Zitting
>            Assignee: Jukka Zitting
>             Fix For: 2.0.0
> Once Apache Tika is released with a resolution to TIKA-175 (making Tika available to
Java 1.4 projects), we should replace our direct parser library dependencies with Tika parsers.
Ideally we'd just use the Tika AutoDetectParser that'll automatically detect the type of a
binary and parse it accordingly, solving JCR-728.
> I guess we should keep some level of backwards compatibility with existing textFilterClasses="..."
configurations, perhaps by keeping the existing TextExtractor classes as wrappers around respective
Tika parsers.

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

View raw message