[opencms-dev] Solr indexing of custom content type
Thorsten Duhn
duhn at regio-gmbh.de
Tue Nov 27 09:24:52 CET 2018
Hello,
thanks a lot for your patience.
The "q=content_de:..." did it finally. In JSP-Search config a query modifier "content_de:%(query)" is needed.
Which in general leaves the question how to do this on multilanguage sites, but for now that's not my problem.
Nevertheless I'm curious. Docs read like "Even with no configuration at all, you have a, thought simple, full text search" (and Apollo template has a quite more difficult search configuration I don't understand).
https://documentation.opencms.org/opencms-documentation/more-opencms-features/solr-search-integration/full-text-search-with-solr/<https://documentation.opencms.org/opencms-documentation/more-opencms-features/solr-search-integration/full-text-search-with-solr/index.html>
Regards,
Thorsten
--
B.Sc. Thorsten Duhn
regio gmbh - Institut für Regionalentwicklung und Informationssysteme
Uhlhornsweg 99a, D-26129 Oldenburg
Fon +49 (0)441 798-2818
Fax +49 (0)441 798-3002
Mail duhn at regio-gmbh.de
Web http://www.regio-gmbh.de
Geschäftsführer: Bernd Kramer, Dr. Uwe Kröcher
Amtsgericht Oldenburg (HRB 3903), USt-IdNr. DE187010339
Von: opencms-dev-bounces at opencms.org [mailto:opencms-dev-bounces at opencms.org] Im Auftrag von Daniel Seidel
Gesendet: Montag, 26. November 2018 11:23
An: opencms-dev at opencms.org
Betreff: Re: [opencms-dev] Solr indexing of custom content type
Hi Torsten,
from your files, it looks like the contents are indexed correctly - the content_de of the containerpage contains "Lorem ipsum" as well. So maybe your query is not correct? Try, e.g., with q=content_de:Lorem&fq=con_locales:de&fq=type:containerpage to just find the container pages.
Best Daniel.
Am 22.11.18 um 11:08 schrieb Thorsten Duhn:
Hello,
not quite sure this is what you mean (I guess sent the first one already), but here they are:
This is the response for the search I'm trying to implement in "handleSolrSelect-jspsearch.xml" (note, containerPageOnly is true for "custom-content" and new "custom-flexible", false for "custom-news" which has detail page).
/handleSolrSelect?q=lorem&fq=type:custom-*&fq=con_locales:*&fq=parent-folders:%22/sites/default/%22
And this is for the page that is not found in "handleSolrSelect-containerpage.xml" containing "custom-flexible" content.
/handleSolrSelect?q=*&fq=parent-folders:%22/sites/default/projekte/dummy/%22
Regards,
Thorsten
--
B.Sc. Thorsten Duhn
regio gmbh - Institut für Regionalentwicklung und Informationssysteme
Uhlhornsweg 99a, D-26129 Oldenburg
Fon +49 (0)441 798-2818
Fax +49 (0)441 798-3002
Mail duhn at regio-gmbh.de<mailto:duhn at regio-gmbh.de>
Web http://www.regio-gmbh.de
Geschäftsführer: Bernd Kramer, Dr. Uwe Kröcher
Amtsgericht Oldenburg (HRB 3903), USt-IdNr. DE187010339
Von: opencms-dev-bounces at opencms.org<mailto:opencms-dev-bounces at opencms.org> [mailto:opencms-dev-bounces at opencms.org] Im Auftrag von Daniel Seidel
Gesendet: Donnerstag, 22. November 2018 08:37
An: opencms-dev at opencms.org<mailto:opencms-dev at opencms.org>
Betreff: Re: [opencms-dev] Solr indexing of custom content type
Hi Thorsten,
please send the Solr-Handler-Response for a) the content b) the container-page the content is on. You can search like q=*:*&fq=path:"/the/root/path/to/the/resource".
Best, Daniel.
Am 22.11.18 um 08:35 schrieb Thorsten Duhn:
Just to add, it's not about nested schemas, a simpler "flexible" content has the same problem for me: I cannot find anything via its container page. So it seems obvious that there is a problem on my side, as of course there are several working usages, even OC doc site. But I guess, I double checked and tested mostly everything I can imagine.
Regards
--
B.Sc. Thorsten Duhn
regio gmbh - Institut für Regionalentwicklung und Informationssysteme
Uhlhornsweg 99a, D-26129 Oldenburg
Fon +49 (0)441 798-2818
Fax +49 (0)441 798-3002
Mail duhn at regio-gmbh.de<mailto:duhn at regio-gmbh.de>
Web http://www.regio-gmbh.de
Geschäftsführer: Bernd Kramer, Dr. Uwe Kröcher
Amtsgericht Oldenburg (HRB 3903), USt-IdNr. DE187010339
Von: opencms-dev-bounces at opencms.org<mailto:opencms-dev-bounces at opencms.org> [mailto:opencms-dev-bounces at opencms.org] Im Auftrag von Thorsten Duhn
Gesendet: Mittwoch, 21. November 2018 09:22
An: The OpenCms mailing list
Betreff: Re: [opencms-dev] Solr indexing of custom content type
Hello,
no, all this does not change anything. "locale-available" already was set correctly to "de", and I cannot spot any difference or problem remaining in formatter config. Attached it, but tried a lot of possible settings in it with no success. The only game change is containerPageOnly in schema, but that is useless for site search where I need to find the container page containing this content.
Sure this really works? I remember there were also problems on this before <cms:search> was introduced. And with OC10 not only <cms:search> was new, also Apollo template came with a simpler "Flexible" content type, with no nested content. But I do not find indication in docs, that nested structure is not valid anymore.
I'm quite helpless for now. Unfortunately official Alkacon support only offers packages of ten incidents, which is way too much for our needs.
I guess I will now try to create a new content type - based on Apollo approach, not nested. But this will break upgrade path for some of our sites, where site search would be necessary as well.
Any further ideas?
Regards,
Thorsten
--
B.Sc. Thorsten Duhn
regio gmbh - Institut für Regionalentwicklung und Informationssysteme
Uhlhornsweg 99a, D-26129 Oldenburg
Fon +49 (0)441 798-2818
Fax +49 (0)441 798-3002
Mail duhn at regio-gmbh.de<mailto:duhn at regio-gmbh.de>
Web http://www.regio-gmbh.de
Geschäftsführer: Bernd Kramer, Dr. Uwe Kröcher
Amtsgericht Oldenburg (HRB 3903), USt-IdNr. DE187010339
Von: opencms-dev-bounces at opencms.org<mailto:opencms-dev-bounces at opencms.org> [mailto:opencms-dev-bounces at opencms.org] Im Auftrag von Daniel Seidel
Gesendet: Dienstag, 20. November 2018 17:14
An: opencms-dev at opencms.org<mailto:opencms-dev at opencms.org>
Betreff: Re: [opencms-dev] Solr indexing of custom content type
Hi Thorsten,
looks like your contents by themself are indexed correctly. You may again have a look at the formatter-configuration xml file, of the formatter you drop your content to the container pages and you may find out if you find containerpages at all and you could try to set the "locale-available" property at the site root to "de" if you use only "de" - and then re-index a page you want to find.
Best, Daniel.
Am 20.11.18 um 15:57 schrieb Thorsten Duhn:
Hello,
I'm not sure about posting files in here, but also the relevant XSD here. Anything wrong with searchsettings?
Regards,
Thorsten
--
B.Sc. Thorsten Duhn
regio gmbh - Institut für Regionalentwicklung und Informationssysteme
Uhlhornsweg 99a, D-26129 Oldenburg
Fon +49 (0)441 798-2818
Fax +49 (0)441 798-3002
Mail duhn at regio-gmbh.de<mailto:duhn at regio-gmbh.de>
Web http://www.regio-gmbh.de
Geschäftsführer: Bernd Kramer, Dr. Uwe Kröcher
Amtsgericht Oldenburg (HRB 3903), USt-IdNr. DE187010339
Von: opencms-dev-bounces at opencms.org<mailto:opencms-dev-bounces at opencms.org> [mailto:opencms-dev-bounces at opencms.org] Im Auftrag von Daniel Seidel
Gesendet: Dienstag, 20. November 2018 15:01
An: opencms-dev at opencms.org<mailto:opencms-dev at opencms.org>
Betreff: Re: [opencms-dev] Solr indexing of custom content type
Hi Thorsten,
when you look at the search result for the content itself, what's in the fields "content_de"/"content_en" ? This is what should be added to the container page.
Best, Daniel.
Am 20.11.18 um 14:53 schrieb Thorsten Duhn:
Hello again,
no, I wasn't right, it is not solved I should have taken a look more carefully.
What changed the situation in my test was the change to containerpageonly="false". With that my custom-content elements were found. But of course the link generated is to its XML file, not to the containing container page. Changing back returns me to the given problem, I get one result where there should be two, only the news detail page.
What is found is text contained in page titles. But not its content. Not even contents title (which obviously isn't part of the nested types).
Quite bizarre.
Regards,
Thorsten
--
B.Sc. Thorsten Duhn
regio gmbh - Institut für Regionalentwicklung und Informationssysteme
Uhlhornsweg 99a, D-26129 Oldenburg
Fon +49 (0)441 798-2818
Fax +49 (0)441 798-3002
Mail duhn at regio-gmbh.de<mailto:duhn at regio-gmbh.de>
Web http://www.regio-gmbh.de
Geschäftsführer: Bernd Kramer, Dr. Uwe Kröcher
Amtsgericht Oldenburg (HRB 3903), USt-IdNr. DE187010339
Von: opencms-dev-bounces at opencms.org<mailto:opencms-dev-bounces at opencms.org> [mailto:opencms-dev-bounces at opencms.org] Im Auftrag von Thorsten Duhn
Gesendet: Dienstag, 20. November 2018 10:29
An: The OpenCms mailing list
Betreff: Re: [opencms-dev] Solr indexing of custom content type
Hello,
wow, thanks a lot, now it works. I guess removing search settings from nested XSD did the trick. So I did too much before.
Thanks,
Thorsten
--
B.Sc. Thorsten Duhn
regio gmbh - Institut für Regionalentwicklung und Informationssysteme
Uhlhornsweg 99a, D-26129 Oldenburg
Fon +49 (0)441 798-2818
Fax +49 (0)441 798-3002
Mail duhn at regio-gmbh.de<mailto:duhn at regio-gmbh.de>
Web http://www.regio-gmbh.de
Geschäftsführer: Bernd Kramer, Dr. Uwe Kröcher
Amtsgericht Oldenburg (HRB 3903), USt-IdNr. DE187010339
Von: opencms-dev-bounces at opencms.org<mailto:opencms-dev-bounces at opencms.org> [mailto:opencms-dev-bounces at opencms.org] Im Auftrag von Daniel Seidel
Gesendet: Dienstag, 20. November 2018 09:59
An: opencms-dev at opencms.org<mailto:opencms-dev at opencms.org>
Betreff: Re: [opencms-dev] Solr indexing of custom content type
Hi Thorsten,
for debugging, I would make the custom-content findable by itself (searchsettings containerpageonly="false") than you only need to touch one of these contents (maybe with rewrite content, I always forget about that) and then use the solr handler to see how it is indexed. Moreover, check, where the searchsettings are done. I think in 10.5 it's necessary to have them in the main XSD.
Best, Daniel.
Am 20.11.18 um 09:49 schrieb Thorsten Duhn:
Hello,
sorry forgot to mention (so many different places involved), of course that is/always was set true.
Regards,
Thorsten
--
B.Sc. Thorsten Duhn
regio gmbh - Institut für Regionalentwicklung und Informationssysteme
Uhlhornsweg 99a, D-26129 Oldenburg
Fon +49 (0)441 798-2818
Fax +49 (0)441 798-3002
Mail duhn at regio-gmbh.de<mailto:duhn at regio-gmbh.de>
Web http://www.regio-gmbh.de
Geschäftsführer: Bernd Kramer, Dr. Uwe Kröcher
Amtsgericht Oldenburg (HRB 3903), USt-IdNr. DE187010339
Von: opencms-dev-bounces at opencms.org<mailto:opencms-dev-bounces at opencms.org> [mailto:opencms-dev-bounces at opencms.org] Im Auftrag von Daniel Seidel
Gesendet: Dienstag, 20. November 2018 09:47
An: opencms-dev at opencms.org<mailto:opencms-dev at opencms.org>
Betreff: Re: [opencms-dev] Solr indexing of custom content type
Hi Thorsten,
you should look at the formatter configuration of your content. There's a boolean option, you have to check if you want to index your content with the page - i think it's called "Search content".
Best, Daniel.
Am 20.11.18 um 09:27 schrieb Thorsten Duhn:
Hello,
I have a "custom-content" resource type, a bit like a evolution of what was "Flexible" in OC8/OC9, with nested choice of HTML, code, image and alike. Also I have a "custom-news" type with no nested content. Having put Lorem ipsum content on my site, a container page with "custom-content" and one item for "custom-news" list/detail. But when I try to get <cms:search> running I only find "custom-news" detail pages, not container pages having "custom-content".
/handleSolrSelect?q=ipsum&fq=type:custom-*+OR+type:containerpage&fq=parent-folders:%22/sites/default/%22
I don't know where else to look at. All schemas have <searchsettings> (for all elements, "searchcontent" true and false), mainly as I see it in OC9 "flexible.xsd". I reindexed on- and offline indexes several times, modified content elements and saved, restarted Tomcat, but always find only one of two elements.
How can I find content from cointainerpages? Where can I find indications of what goes wrong?
System is OC 10.5.4, content is German (and "fq" of above request contains "con_locales:de").
Regards,
Thorsten
--
B.Sc. Thorsten Duhn
regio gmbh - Institut für Regionalentwicklung und Informationssysteme
Uhlhornsweg 99a, D-26129 Oldenburg
Fon +49 (0)441 798-2818
Fax +49 (0)441 798-3002
Mail duhn at regio-gmbh.de<mailto:duhn at regio-gmbh.de>
Web http://www.regio-gmbh.de
Geschäftsführer: Bernd Kramer, Dr. Uwe Kröcher
Amtsgericht Oldenburg (HRB 3903), USt-IdNr. DE187010339
_______________________________________________
This mail is sent to you from the opencms-dev mailing list
To change your list options, or to unsubscribe from the list, please visit
http://lists.opencms.org/cgi-bin/mailman/listinfo/opencms-dev
--
Kind Regards,
Daniel.
-------------------
Daniel Seidel
Alkacon Software GmbH & Co. KG - The OpenCms Experts
http://www.alkacon.com - http://www.opencms.org
_______________________________________________
This mail is sent to you from the opencms-dev mailing list
To change your list options, or to unsubscribe from the list, please visit
http://lists.opencms.org/cgi-bin/mailman/listinfo/opencms-dev
--
Kind Regards,
Daniel.
-------------------
Daniel Seidel
Alkacon Software GmbH & Co. KG - The OpenCms Experts
http://www.alkacon.com - http://www.opencms.org
_______________________________________________
This mail is sent to you from the opencms-dev mailing list
To change your list options, or to unsubscribe from the list, please visit
http://lists.opencms.org/cgi-bin/mailman/listinfo/opencms-dev
--
Kind Regards,
Daniel.
-------------------
Daniel Seidel
Alkacon Software GmbH & Co. KG - The OpenCms Experts
http://www.alkacon.com - http://www.opencms.org
_______________________________________________
This mail is sent to you from the opencms-dev mailing list
To change your list options, or to unsubscribe from the list, please visit
http://lists.opencms.org/cgi-bin/mailman/listinfo/opencms-dev
--
Kind Regards,
Daniel.
-------------------
Daniel Seidel
Alkacon Software GmbH & Co. KG - The OpenCms Experts
http://www.alkacon.com - http://www.opencms.org
_______________________________________________
This mail is sent to you from the opencms-dev mailing list
To change your list options, or to unsubscribe from the list, please visit
http://lists.opencms.org/cgi-bin/mailman/listinfo/opencms-dev
--
Kind Regards,
Daniel.
-------------------
Daniel Seidel
Alkacon Software GmbH & Co. KG - The OpenCms Experts
http://www.alkacon.com - http://www.opencms.org
_______________________________________________
This mail is sent to you from the opencms-dev mailing list
To change your list options, or to unsubscribe from the list, please visit
http://lists.opencms.org/cgi-bin/mailman/listinfo/opencms-dev
--
Kind Regards,
Daniel.
-------------------
Daniel Seidel
Alkacon Software GmbH & Co. KG - The OpenCms Experts
http://www.alkacon.com - http://www.opencms.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://webmail.opencms.org/pipermail/opencms-dev/attachments/20181127/222c37f3/attachment.htm>
More information about the opencms-dev
mailing list