Proposal for lightweight semantification using semantify³

[WolfgangFahl]

A query such as https://github.com/WDscholia/scholia/blob/main/scholia/app/templates/event-series-index_list.sparql

SELECT
  ?events ?short_name
  ?event_series ?event_seriesLabel ?event_seriesDescription
  (CONCAT("/event-series/", SUBSTR(STR(?event_series), 32)) AS ?event_seriesUrl)
WITH {
  SELECT (COUNT(?event) AS ?events) ?event_series {  
    { ?event_series wdt:P31 wd:Q47258130 . }
    UNION 
    { ?event_series wdt:P31 wd:Q47459256 . }
    OPTIONAL { ?event_series ^wdt:P179 ?event }
  }
  GROUP BY ?event_series
} AS %events
WHERE {
  INCLUDE %events
  OPTIONAL { ?event_series wdt:P1813 ?short_name }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],mul,en,da,de,fr". }
}
ORDER BY DESC(?events)

is currently semantified by snapquery embedding the meta-data in yaml file

- name: event-series-index_list
  namespace: named_queries
  domain: scholia.toolforge.org
  query_id: [email protected]
  sparql: "SELECT\n  ?events ?short_name\n  ?event_series ?event_seriesLabel ?event_seriesDescription\n\
    \  (CONCAT(\"/event-series/\", SUBSTR(STR(?event_series), 32)) AS ?event_seriesUrl)\n\
    WITH {\n  SELECT (COUNT(?event) AS ?events) ?event_series {  \n    { ?event_series\
    \ wdt:P31 wd:Q47258130 . }\n    UNION \n    { ?event_series wdt:P31 wd:Q47459256\
    \ . }\n    OPTIONAL { ?event_series ^wdt:P179 ?event }\n  }\n  GROUP BY ?event_series\n\
    } AS %events\nWHERE {\n  INCLUDE %events\n  OPTIONAL { ?event_series wdt:P1813\
    \ ?short_name }\n  SERVICE wikibase:label { bd:serviceParam wikibase:language\
    \ \"[AUTO_LANGUAGE],mul,en,da,de,fr\". }\n}\nORDER BY DESC(?events)\n"
  url: https://raw.githubusercontent.com/WDscholia/scholia/main/scholia/app/templates/event-series-index_list.sparql
  title: event-series-index_list
  description: event-series-index_list
  comment: ''

which then needs to be parsed and may be converted to RDF format as proposed in snapquery issue #59 Snap query should consider using the sparql-examples style of encoding queries as their own entities as the basis of it's data interchange.

Instead of having Snapquery try to collect meta-data from all queries it would IMHO be much more straightforward to have the meta - data annoation right in the queries.

The https://github.com/BITPlan/semantify3 semantify³ project proposes an annotation for that:
the yaml snippet is simply embedded in the SPARQL comment

#```yaml
## 🌐🕸
#query:
#  url: https://raw.githubusercontent.com/WDscholia/scholia/main/scholia/app/templates/event-series-index_list.sparql
# title: event-series-index_list
#  description: event-series-index_list
#  comment: ''
#  name: event-series-index_list
#  namespace: named_queries
#  domain: scholia.toolforge.org
#  query_id: [email protected]
# ```

semantify³ will then happily collect the metadata from the files. This process is IMHO much more straightforward then the current github scraping procedure as implemented in https://github.com/WolfgangFahl/snapquery/blob/main/snapquery/scholia.py

and also the snapquery solution can be made compatible to respect this metadata.

Instead of the current approach to load an SQL database. the triple conversion is readily available. To show off this capability i propose to add a SPARQL endpoint based on Vincent Emonet's https://github.com/vemonet/rdflib-endpoint

that will allow to do queries about queries and queries about endpoints and of course queries about runtime behavior.

going this lightweight route first before adding these features in an heavyweight style to snapquery seems to be more adequated for the current pressing need to get queries in the migration phase compared and analyzed quickly.