manifoldcf-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bisonti Mario <Mario.Biso...@vimar.com>
Subject R: How to set Tika with ManifoldCF and Solr
Date Thu, 11 Oct 2018 12:37:22 GMT
This is my solr output connection:

[cid:image002.jpg@01D4616F.EA54D800]

I tried to put content_type as “Mime type field name:” but the result is always the same

Could be that, unchecking the flag, ManifoldCF doesn’t use the mime types specified?

I am using a snapshot version of ManifoldCF of three monts  ago.




Da: Karl Wright <daddywri@gmail.com>
Inviato: giovedì 11 ottobre 2018 14:20
A: user@manifoldcf.apache.org
Oggetto: Re: How to set Tika with ManifoldCF and Solr

I confirmed that both the Tika Service transformer and the Tika transformer check the same
exact mime type:

>>>>>>
  @Override
  public boolean checkMimeTypeIndexable(VersionContext pipelineDescription, String mimeType,
IOutputCheckActivity checkActivity)
    throws ManifoldCFException, ServiceInterruption
  {
    // We should see what Tika will transform
    // MHL
    // Do a downstream check
    return checkActivity.checkMimeTypeIndexable("text/plain;charset=utf-8");
  }
<<<<<<

So: please verify that your Solr connection is set up correctly and the "use extracting update
handler" box is UNCHECKED.

Thanks,
Karl


On Thu, Oct 11, 2018 at 8:16 AM Karl Wright <daddywri@gmail.com<mailto:daddywri@gmail.com>>
wrote:
When you uncheck the "use extracting update handler" checkbox, the Solr connection only accepts
text/plain, and no binary formats.  The Tika extractor, though, should set the mime type always
to "text/plain".  Since the Simple History says otherwise, I wonder if there's a problem with
the external Tika extractor.  Perhaps you can try the internal one to get your pipeline working
first?  If the external one does not send the right mime type, then we need to correct that
so you should open a ticket.

Thanks,
Karl


On Thu, Oct 11, 2018 at 8:10 AM Bisonti Mario <Mario.Bisonti@vimar.com<mailto:Mario.Bisonti@vimar.com>>
wrote:
Now the document isn’t ingested by solr because I obtain:

Solr connector rejected document due to mime type restrictions: (application/vnd.openxmlformats-officedocument.spreadsheetml.sheet)


But the mime type is on the tab


And the settings worked well when I used Tika inside solr.

Could you help me?
Thanks

Da: Bisonti Mario <Mario.Bisonti@vimar.com<mailto:Mario.Bisonti@vimar.com>>
Inviato: giovedì 11 ottobre 2018 14:03
A: user@manifoldcf.apache.org<mailto:user@manifoldcf.apache.org>
Oggetto: R: How to set Tika with ManifoldCF and Solr


My mistake…
As you wrote me I had to uncheck “use extracting update handler”

Now I have to understand the field mentioned in schema etc.

Da: Bisonti Mario <Mario.Bisonti@vimar.com<mailto:Mario.Bisonti@vimar.com>>
Inviato: giovedì 11 ottobre 2018 13:45
A: user@manifoldcf.apache.org<mailto:user@manifoldcf.apache.org>
Oggetto: R: How to set Tika with ManifoldCF and Solr

I see the job processed but without the document inside.
10-11-2018 13:32:25.649

job end

1539153700219(G_IT_Area_condivisa_Mario_XLSM)

0

1

10-11-2018 13:32:14.211

job start

1539153700219(G_IT_Area_condivisa_Mario_XLSM)

0

1





Have I to uncheck, on my Solr output connection the “Use the Extract Update Handler”?






Da: Karl Wright <daddywri@gmail.com<mailto:daddywri@gmail.com>>
Inviato: giovedì 11 ottobre 2018 13:36
A: user@manifoldcf.apache.org<mailto:user@manifoldcf.apache.org>
Oggetto: Re: How to set Tika with ManifoldCF and Solr

Please have a look at your "Simple History" report to see why the documents aren't getting
indexed.

Thanks,
Karl


On Thu, Oct 11, 2018 at 7:10 AM Bisonti Mario <Mario.Bisonti@vimar.com<mailto:Mario.Bisonti@vimar.com>>
wrote:
Thanks Karl.
I tried, but it doesn’t index documents.
It seemes that it doesn’t see them?

Perhaps is the “Ignore Tika exception that I don’t know where to set in ManifoldCF  the
problem?





Da: Karl Wright <daddywri@gmail.com<mailto:daddywri@gmail.com>>
Inviato: giovedì 11 ottobre 2018 12:24
A: user@manifoldcf.apache.org<mailto:user@manifoldcf.apache.org>
Oggetto: Re: How to set Tika with ManifoldCF and Solr

Hi Mario,

(1) When you use the Tika server externally, you do not get the boilerpipe HTML extractor
available for configuration and use.  That is because it's external now.
(2) In your Solr connection, you want to uncheck the box that says "use extracting update
handler", and you want to change the output handler from "/update/extract" to just "/update".

Karl


On Thu, Oct 11, 2018 at 4:45 AM Bisonti Mario <Mario.Bisonti@vimar.com<mailto:Mario.Bisonti@vimar.com>>
wrote:
Hallo.
I would like to use Tika server started from command line into ManifoldCF so, ManifoldCF as
Trasformation connector, process with Tika and index to the output connecto Solr.

I started Tika server:
java -jar /opt/tika/tika-server-1.19.1.jar

After, I created a transformation connection with TikaServer: localhost and Tika port 998
and connection works.

After, I created a job and in the Tab Connection I inserted the Transformation yet created
Before the Output Solr.



Note that I don’t see the tab “Excepition” and “Boilerplate”
Why this?

Furthermore, if I start the job, I see that Solr hangs with exception:
2018-10-11 10:03:47.268 WARN  (qtp1223240796-17) [   x:core_share] o.e.j.s.HttpChannel /solr/core_share/update/extract
java.lang.NoClassDefFoundError: org/apache/tika/exception/TikaException
        at java.lang.Class.forName0(Native Method) ~[?:?]
        at java.lang.Class.forName(Class.java:374) ~[?:?]

infact, I renamed the tika .jar:
in the folder : solr/contrib/extraction/lib to be sure that solr doesn’t use Tika because
I would like that Manifoldcfuses Tika buti t doesn’t work.

Have I to configure solr to don’t use Tika I suppose.

How to do this?

I see https://datafari.atlassian.net/wiki/spaces/DATAFARI/pages/107708451/Data+Extraction+Tika+Embedded+in+Solr+Deactivation+Configuration<https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fdatafari.atlassian.net%2Fwiki%2Fspaces%2FDATAFARI%2Fpages%2F107708451%2FData%2BExtraction%2BTika%2BEmbedded%2Bin%2BSolr%2BDeactivation%2BConfiguration&data=01%7C01%7CMario.Bisonti%40vimar.com%7C5e0dce2092dd4af370de08d62f73f871%7Ca1f008bcd59b4c668f8760fd9af15c7f%7C1&sdata=KGTBcRjlMQIfrf1gVfyWWZlrFnaEBsl6oW5VP5Jq%2Ba8%3D&reserved=0>
but I haven’t Datafari, so, in a Solr standard configuration, how could I deactivated the
tika ?

Thanks a lot

Mario

Mime
View raw message