Hi,
Morphlines stuff is not needed at all. This is a Mapreduce/Hadoop integration of Solr (see documentation) – mostly command line tools around Solr and Hadoop. FYI: In Solr we don’t show the warnings, because otherwise the user would get a lot of useless warnings. We may fix this when TIKA 2.0 comes with the tika-parsers module split into multiple parts. In Solr we currently removed all TIKA dependencies that conflict with the rest of Solr/Lucene or are not useful for fulltext indexing. We only left those that are useful for “fulltext” extraction (e.g. office document formats). But we have no parsers for CLASS files (breaks, because of ASM conflict) or Netcdf files (License issues in older versions of TIKA). I see no reason to show these warnings, because if you use Solr as documented, it should work correctly. We no longer support running Solr inside foreign Application Servers. So everything should work out of box. Uwe ----- Uwe Schindler H.-H.-Meier-Allee 63, D-28213 Bremen <http://www.thetaphi.de/> http://www.thetaphi.de eMail: u...@thetaphi.de From: Steven White [mailto:swhite4...@gmail.com] Sent: Wednesday, February 03, 2016 5:44 PM To: user@tika.apache.org Subject: Re: Using Tika that comes with Solr 5.2 Nick, that would be a good think to do: changing Ignore to Warn otherwise newcomers will have no clue why this isn't working. Another question to the team regarding this topic. I see JARs under solr\contrib\morphlines-cell\lib\ and solr\contrib\morphlines-core\lib\ The ones under "morphlines-cell" there are 2 files with "tika" as their name. My question is, do I need those for general Tika usage? The README.txt clearly states "*Experimental*" but doesn't say if I need them to use Tika. Steve On Wed, Feb 3, 2016 at 9:16 AM, Nick Burch <apa...@gagravarr.org <mailto:apa...@gagravarr.org> > wrote: On Wed, 3 Feb 2016, Uwe Schindler wrote: The reason for this behaviour is part of TIKA: If a parser cannot load because of classes it refers to are missing, it is automatically disabled. Because you missed the actual PDF/Powerpoint/… classes, this is what happens for all those parsers. I wonder if it might be worth SOLR changing their default Tika config from Ignore to Warn, so that SOLR users (who probably aren't as clued up on how it all works as the average Tika user) will get to find out more quickly that they've missed something? Nick
