Hello @dipso-num4sci , have you tried setting the skosmos:mainConceptScheme configuration option for the vocabulary?

Thank you Osma.
Yes we've tried to use skosmos:mainConceptScheme.

But finally, we solved this problem by modifying the data itself.
Thank you again.

We re-open this issue because the problem is not really resolved.

We have uploaded this file: https://nextcloud.inrae.fr/s/MPdTs2DLzx9TnFM in Skosmos version 2.16
And set the options like this :

• skosmos:mainConceptScheme = http://opendata.inrae.fr/thesaurusINRAE/thesaurusINRAE
• skosmos:conceptSchemesInHierarchy = false
  But we still have the other ConceptSchemes shown at the second level, under the domain; e.g. ENV environment and natural resources under 01. ENVIRONMENT
  What we would like is to show only the hierarchy which is under INRAE thesaurus (this hierarchy contains only skos:Concept, including some with the same name as our microthesauri)

We have tried to change the data, to change the skosmos configuration... but could not find what is wrong with us.
Any idea of what we could do?

Reopened due to requests above

Hi @tfrancart,
would you have an idea on how we can neutralize the display of the microthesauri in the hierarchy? We think it is the feature you developed for Irstea that is concerned (build the hierarchy from domains/dc:subjects and microthesauri/conceptSchemes). It looks like the option skosmos:mainConceptScheme does not work in this case and we can't find why.

Sorry, it is hard for me to help you here. I don't even have a Skosmos test/dev environment to look at.
Seems to me that this line :

Have you tried cutting the dcterms:subject links from the MT to the domain in the data ? this relation between ConceptScheme is interpreted on the javascript client side if I remember properly and I don't think there is a way to change this.

Thanks Thomas for the tips. I can answer on this:

Have you tried cutting the dcterms:subject links from the MT to the domain in the data ? this relation between ConceptScheme is interpreted on the javascript client side if I remember properly and I don't think there is a way to change this.

Yes, we have tried to remove "dcterms:subject" but skosmos still displays all the ConceptSchemes in the hierarchy.

Absolutely. This version of the file contains the dctems:subject. We had tried with this one : https://nextcloud.inrae.fr/s/6ygRjbWNQ8rGPrJ

(EDIT: fixed the link -Osma)

Dear @osma and @tfrancart ,

Thank you for the various messages we have shared together over the past few months regarding this problem.

Unfortunately, this issue is not resolved, although we have tested many configuration options.
We currently believe that the only way to resolve the issue should be an update of the Skosmos source code, disabling some features developed by Thomas over the last few years.
In fact, these features were specifically created to meet certain Irstea needs, and are therefore no longer necessary.

Do you think it would be possible to move forward together in this direction, by integrating our request into a future version of Skosmos ?
An engineer is currently in charge of the subject and would be available to move forward with you.

Many thanks !

I close this issue; unresolved, but reproduced in #1535 with higher priority.
Thanks !

Reopening; let's keep using this issue instead of opening new ones.

Sorry for the long delays. We have been prioritizing Skosmos 3 development, and the hierarchy component is going to be implemented in a different way in the new front-end.

I think I see the problem. I have tested with the thesaurus data provided in #1396 (comment) and I can replicate the problem locally. Basically, there is currently no code on the front-end side that would ensure that only the default/main concept scheme is displayed at the top level in the hierachy.

I'll see if this can be easily implemented for Skosmos 2.x since we are still doing some maintenance and possibly one more release before retiring 2.x.

Proposed fix in PR #1540. It would be great if you could review this @saubin78 and @dipso-num4sci ! Even better if you can test the code in the PR locally.