odeuropa / knowledge-graph Goto Github PK

The graph http://data.odeuropa.eu/pulse has its label set to "ECCO" instead of "Pulse", as seen in this file:

With the update of philologic, URLs such as

https://artflsrv03.uchicago.edu/philologic4/newberryfrc1217/navigate/28921/1/

should be changed into

https://artflsrv04.uchicago.edu/philologic4.7/newberryfrc1217/navigate/28921/1/

On the Explorer, the list of excerpts is randomly ordered and isn't sorted by the position of the excerpt in the full document, which can be confusing.

This is because the KG doesn't have a notion of position for each fragment.

Proposal: adding a numeric property which tells us at which position in the document the excerpt was taken from.

Related to Odeuropa/explorer#44

Dates formatted like "23XX" will have an incorrectly infered beginning/end dates (2301/2399).

Example: http://data.odeuropa.eu/time/f8104077-b9e4-501d-b89f-67e4c8e11635

When there is not clear parsed time for a smell event, use the document (image or text) time, marking it as "inferred" using RDF-star

It would be interesting to include somehow the topic (e.g. schema:about) of the smell, e.g. war, odour of sanctity...

Keep track in the graph of the words used for extraction, in order to visualise them

We need to work on the relationship between Europeana content and EOKG in both ways:

1. Importing Europeana content in the EOKG with useful metadata. The first stop shop is the APIs that Europeana is providing, https://pro.europeana.eu/page/apis ... but we need guidance from the Historians to know what content is interesting to pull from / cc @Ileemans @williamtullett
2. Share the EOKG data in Europeana. The guide is available at https://pro.europeana.eu/share-your-data/process. Ideally, we would try to get funding to become and develop a Europeana aggregator

Interesting finding: https://odeuropa.tools.eurecom.fr/smells/07b55dd0-45a5-57c4-b763-485766435cb9 … but why “butter” is the smell emission in this example? It got confused with “cigar butts”.

Example: http://data.odeuropa.eu/smell/00100987-a7f9-57f3-a784-087226aeb04c is referenced by 2 different images

The following query should be executed after creating the GraphDB repository. It is used to initialize the Lucene connector for full-text search queries.

PREFIX :<http://www.ontotext.com/connectors/lucene#>
PREFIX inst:<http://www.ontotext.com/connectors/lucene/instance#>
INSERT DATA {
	inst:search :createConnector '''
{
  "fields": [
    {
      "fieldName": "source_value",
      "propertyChain": [
        "http://erlangen-crm.org/current/P67i_is_referred_to_by",
        "http://www.w3.org/1999/02/22-rdf-syntax-ns#value"
      ],
      "indexed": true,
      "stored": true,
      "analyzed": true,
      "multivalued": true,
      "ignoreInvalidValues": false,
      "facet": false
    },
    {
      "fieldName": "source_label",
      "propertyChain": [
        "http://erlangen-crm.org/current/P67i_is_referred_to_by",
        "http://www.w3.org/2000/01/rdf-schema#label"
      ],
      "indexed": true,
      "stored": true,
      "analyzed": true,
      "multivalued": true,
      "ignoreInvalidValues": false,
      "facet": false
    }
  ],
  "languages": [],
  "types": [
    "http://data.odeuropa.eu/ontology/L11_Smell"
  ],
  "readonly": false,
  "detectFields": false,
  "importGraph": false,
  "skipInitialIndexing": false,
  "boostProperties": [],
  "stripMarkup": false
}
''' .
}

1. First solution: creating a script which automates the insertion of the query (maybe in scripts/create_repo.py?)
2. Second solution: updating the README with steps to manually execute the query