search

ArtResearch / solr-docker

0 stars 0 forks source link

SOLR Indexing #2

Closed MinadakisNikos closed 3 years ago

MinadakisNikos commented 3 years ago

Create SOLR indexings on rdfs:label but investigate how to link it with the URI of the main entity (eg bibliography linked to E22)

lklic commented 3 years ago

I think we need to do this through the named graph in which the data resides.

The issue will be with labels that live in the named graphs of vocabularies, we should see if this is a user requirement (i.e. if user want to use full text search to query "bronze sculpture"

lklic commented 3 years ago

we should also look into an external solr index: https://github.com/blazegraph/database/wiki/ExternalFullTextSearch

We may be able to configure the results to return the URI of the entity that we want...

MinadakisNikos commented 3 years ago

adding the actual java interface here: https://blazegraph.com/database/apidocs/com/bigdata/service/fts/FTS.html

instructions on installing here: https://github.com/blazegraph/database/wiki/SOLR_External_Fulltext_Search

@lklic have we installed solr or should we go forward with this. I tried to access it here: http://34.252.34.240:8983/solr/blazegraph/select but it does not exist. I kind of remember that you had installed it, but maybe i am wrong

lklic commented 3 years ago

It isn't installed....

The first step would be to dockerize this, otherwise it will become unmanageable, then we need to figure out how to link the labels of related entities to E22/E21, etc.

building the SOLR index is a very CPU heavy task, so we should probably test this on the new server first, once I finish migration.

FYI I have a duplicate issue here: https://github.com/ArtResearch/pharos-dev/issues/91

lklic commented 3 years ago

I have migrated this issue from the other repo, and I am closing it for now, we should create individual issues for each task related to the SOLR index