URI pattern for Procedure & Lot

[cristianvasquez]

Hello,

Building on the discussion about how URIs should be minted in these mappings, I write a bit about the relevance of the URI schemes, and then provide an alternative

Preamble

The current eForms RML mappings use the URI scheme {ns}id_{notice-id}_{concept}_{trailer}, where:

• {ns} is a base namespace, in this case http://data.europa.eu/a4g/resource/
• {concept} is either (i) an ontology fragment label or (ii) source element label, with a suffix or prefix
• {trailer} is either (i) an ID value (if the resource has one) or (ii) an online computed, deterministic hash

The rationale behind this is explained at: #22

Problem

The choice of URI patterns defines how the shared entities across notices are identified, and as a consequence, the query results.

Example: Currently, if we transform a Notice with the ID: IDNOTICE1, referring to a procedure c30d15d5. We get:

epd:IDNOTICE1 epo:refersToProcedure epd:IDNOTICE1_Procedure_iNVWPykMuHh9p4AY3uzvA2.

epd:IDNOTICE1_Procedure_iNVWPykMuHh9p4AY3uzvA2 a epo:Procedure;
  adms:identifier [
				  a adms:Identifier;
				  epo:hasScheme "some-scheme";
				  skos:notation "c30d15d5" .
  ]
.

If we later transform IDNOTICE2, which refers to the same procedure, the triples will include a new instance of the procedure:

epd:IDNOTICE2 epo:refersToProcedure epd:IDNOTICE2_Procedure_cMBzh6pL4JRkiaCd4DWjT2.

epd:IDNOTICE2_Procedure_cMBzh6pL4JRkiaCd4DWjT2 a epo:Procedure;
  adms:identifier [
				  a adms:Identifier;
				  epo:hasScheme "some-scheme";
				  skos:notation "c30d15d5" .
  ]
.

The following query:

PREFIX epo: <http://data.europa.eu/a4g/ontology#> 

SELECT COUNT distinct ?s WHERE {
	?s a epo:Procedure
}

will yield 2. (One can argue that adms:identifier can be used to identify the procedures as the same, but it's sub-optimal)

Having a URI pattern for procedures which is independent of Notice, such as {some-scheme}-{procedure-id}, will yield triples like:

epd:procedure-c30d15d5 a epo:Procedure.
epd:IDNOTICE1 epo:refersToProcedure epd:procedure-c30d15d5.
epd:IDNOTICE2 epo:refersToProcedure epd:procedure-c30d15d5.

The count of distinct epo:Procedure instances would yield 1 in this case.

Different choices of URI patterns will result in different results.

I wrote examples of alternative URIs for procedures and lots.

• Current
  • http://data.europa.eu/a4g/resource/id_2edebde2-afd7-4adb-bbbb-c569a8906162_Procedure_6dcsBnuV4FTNoRpTZHckqN
• Modified
  • A procedure is independent of notices
  • http://data.europa.eu/a4g/resource/procedure/06dff772-872e-4f5c-9b8b-e7d070eee0af

Example of lot URI.

• Current
  • http://data.europa.eu/a4g/resource/id_2edebde2-afd7-4adb-bbbb-c569a8906162_Lot_LOT-0000
• Modified
  • A lot is part of a procedure, independent of notice.
  • http://data.europa.eu/a4g/resource/procedure/06dff772-872e-4f5c-9b8b-e7d070eee0af/lot/LOT-0000

The PoC works by rewriting URIs through SPARQL queries+code.

Please comment if there are nuances I'm missing