global semantic resolver for electrical data (I Have a Dream)

[VladimirAlexiev]

(Spawned from #98)

To facilitate semantic resolution of electricity-related URLs, ideally we would have a single harmonized namespace, and a global resolver (operated eg by ENTSO-E) that redirects (delegates) to specific authorities (TSO or other issuers of CIM data).

Such strategies are already used for:

• Publishing: DOIs: eg https://dx.doi.org/10.13140/RG.2.2.33957.10729 .
  The delegation is done based on DOI prefix, see P1662 on Wikidata.
  Eg 10.13140 above designates DataCite (which serves as registrant on behalf of ResearchGate).
  The DOI foundation also has registrant pages, eg https://dx.doi.org/10.13140.
• IP addresses (IANA and national agencies like ARIN for US, CA).
  IPs are a fixed number of bits and don't include a separator, so bigger registrants get shorter prefixes and are allocated bigger network blocks.
• Logistics: GS1 Digital Links based on identifiers like GTIN, GLN, GIAI etc: eg https://id.gs1.org/gtin/9506000134352?linkType=all .
  The delegation is done based on GS1 Company Prefix, see P3193 on Wikidata.
  Eg 95060001343 above is one of the prefixes for "GS1 Global Office" (search at https://www.gs1us.org/tools/gs1-company-database-gepir).
  GTIN etc are fixed-width, so similarly to IP addresses, bigger registrants get shorter prefixes.
• Legal body identifiers: GLEIF delegates LEI issuance to LOUs. But rather than delegating the resolution, it keeps a central LEI database that is synchronized often with LOU data.

What is the situation in electricity:

ENTSO-E has defined Energy Identification Code (EIC, eg see P8645 on Wikidata and Energy Identification Code on wikipedia) and delegates its issuance to LIOs. As of 2024-12-17, there are 62 LIOs3

• It has a central database:
  • https://www.entsoe.eu/data/energy-identification-codes-eic/eic-approved-codes/: lists 59.5k (select "EIC type: All codes" and "Show: 100")
  • https://transparency.ontotext.com/graphdb/sparql (ask for login): lists 58/6k with this SPARQL query

PREFIX tr: <https://transparency.ontotext.com/resource/tr/>
select (count(*) as ?c) {?x tr:eic ?y} 

• However, the central database doesn't offer individually resolvable pages per EIC. Furthermore, it is incomplete.
• The Wikidata page lists over 20 "source website for the property", and eg for https://iodb.elia.be/en/publiceiclist/indexpartialview the first two 22T20160907---1Y, 22T20160907----1 are missing. These are "Internal Lines" but other Internal Lines are present in the central database
• ENTSO-G https://www.entsog.eu/approved-codes refers to the ENTSO-E list, but my impression is that Gas-related codes are more sparsely represented in the central database
• The first two EIC chars are the LIO code so potentially could be used for redirecting.
  However, LIOs don't have an obligation to publish resolvable pages; https://eepublicdownloads.entsoe.eu/clean-documents/EDI/Library/EIC_Reference_Manual_Release_5_5.pdf#page=8.74 mandates:
  • to provide a local registry on a web-page accessible by third parties. The LIO publishes the energy sector (gas or electricity or both) and the EIC code types that it covers;
  • to publish the list of all EIC codes allocated by the LIO in a processable form according to the EIC implementation guide that can be downloaded by third party

CIM/CGMES requires that node URNs are based on UUIDs.

• EICs of major resources are recorded as additional identifiers
• but the main (canonical) mRID and the URN are UUIDs
• UUIDs don't have a prefix that would allow partitioning/delegation, so we need to rely on some URL prefix before the UUID

ENTSO-E could still operate a central resolver and redirect per prefix, eg
https://energy.referencedata.eu/resource/statnett/<uuid> vs https://energy.referencedata.eu/resource/elia/<uuid>

• This is a bit better than relying on TSO-specific namespaces, eg
  https://data.statnett.no/<uuid> vs https://iodb.elia.be/<uuid>
• Of course, it's still up to authorities to serve meaningful data for these URLs

Note: https://energy.referencedata.eu/ currently has a number of important lists, but no per-resource resolution.
It serves whole semantic files, eg https://energy.referencedata.eu/StandardReasonCodeTypeList.ttl
Needless to say, the EIC register https://energy.referencedata.eu/EIC.ttl is pretty huge.