The system should be able to provide search results based on semantic information

[javadch]

The semantic search should be integrated into the current search feature in the most transparent way. However, the search results should be clear about the reason and root of their participation in the search results. In order to be able to perform the semantic search, the following work items should be done:

• [ ] A mechanism to import an ontology into the system should be available. This import facility can obtain the chosen ontology via (any of) file, web services, or SPARQL query methods.
• [ ] A mechanism and logic should be in place to allow data managers to upgrade a currently in use ontology. The next version of the ontology may contain deprecated, replaced or even deleted IRIs.
• [ ] During and after defining a data structure as well as during and after uploading a dataset, there should be an easy and interactive way to allow a user to annotate the structure (of the dataset).
• [ ] There should be a tool that detects and reports any possible inconsistency in annotations e.g. when a single variable is linked to different annotations (terms) by different users. Such a tool should provide features to correct the mistake.
• [ ] The semantic search should be integrated into the normal search as transparent as possible, provided that the origin of search result items and why they are part of the result is communicated to the user.
• [ ] If no ontology is registered in the system or based on the system administrator's decision, it should be possible to turn off the semantic search without affecting the conventional search feature.

[jmkeil]

For the ontology import, you might want to use SPARQL Generate to import terminologies from webservices or other sources.