[
https://issues.apache.org/jira/browse/TIKA-1867?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16005013#comment-16005013
]
Daniel Conn commented on TIKA-1867:
-----------------------------------
Hi [~gagravarr],
After trying the link earlier I have tried to exclude this parser but it still
seems to be calling it and in turn trying to call ffmpeg and exiftools due to
the tika-external-parsers.xml file. I too am looking for a solution to this,
perhaps one check if these things exist on startup and then cache this, instead
of many checks for the same program? Or a TikaConfig constructor which allows
you to explicitly remove constructors? These are just ideas though!
Just in case I've got the wrong end of the stick here is what I put in the
config file. Could you kindly confirm this was correct, or where I'm going
wrong:
<?xml version="1.0" encoding="UTF-8"?>
<properties>
<parsers>
<parser class="org.apache.tika.parser.DefaultParser">
<parser-exclude
class="org.apache.tika.parser.external.CompositeExternalParser"/>
</parser>
</parsers>
</properties>
Thanks and Kind Regards
Dan
> Tika external parsers cannot be turned off without patching the
> tika-app-XX.jar
> -------------------------------------------------------------------------------
>
> Key: TIKA-1867
> URL: https://issues.apache.org/jira/browse/TIKA-1867
> Project: Tika
> Issue Type: Bug
> Components: parser
> Affects Versions: 1.11
> Reporter: Roman Kratochvil
>
> The CompositeExternalParser calls ExternalParsersFactory.create() which
> always uses configuration from
> org/apache/tika/parser/external/tika-external-parsers.xml. The issue is that
> this introduces performance regression as the parser initialization checks
> for presence of external commands (ffmpeg, exiftool) and that takes time.
> Unfortunately, there is no way how to turn off this functionality without
> patching the tika-app JAR -- one has to either change the
> tika-external-parsers.xml or remove the whole CompositeExternalParser from
> list of services in /META-INF/services/org.apache.tika.parser.Parser.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
