fixes needed to XML/WithFullCombinationOfFields

[VladimirAlexiev]

@sboeckelmann @dakbhavesh I converted XML/WithFullCombinationOfFields to JSON using your tool and then had to make the following fixes (you can see most of them in https://github.com/VladimirAlexiev/EPCIS/commit/47ad32804186eaab4f946cb8a47c65701d74400f):

• remove prefix "cbvmda": "urn:epcglobal:cbv:mda" since it's declared in the EPCIS context and I get error "protected term redefinition"
• redefine "ext1": "http://ext.com/ext1/" to have a trailing slash (same for ext2, ext3) since namespaces should end at a reasonable delimiter
• change "someFurtherMetadata" to "ext1:someFurtherMetadata" because that's not a valid EPCIS field

Similarly in Example-TransactionEvent-2020_07_03y.xml:

• redefine xmlns:rail="urn:gs1:epcisapp:rail:" to have a trailing colon.
• (I also added a reasonable nested context but you can't do that in XML)

  "@context": [
  "https://gs1.github.io/EPCIS/epcis-context-simple.jsonld",
  {"rail": "urn:gs1:epcisapp:rail:",
   "rail:vehicleMasterGIAI": {"@type": "@id"},
   "rail:trainAxleCount": {"@type": "xsd:integer"},
   "rail:vehicleAxleCount": {"@type": "xsd:integer"},
   "rail:vehiclePosition": {"@type": "xsd:integer"},
   "rail:vehicleUniquelyIdentified": {"@type": "xsd:boolean"}
  }
  ]

Please make them in the XML. Thanks! (And quickly close this issue lest Craig slaps us :-)

[CraigRe]

@sboeckelmann , @dakbhavesh , @VladimirAlexiev ,

• I added a trailing slash to ext1, ext2, ext3
• all instances of "someFurtherMetadata" already appeared to be "ext1:someFurtherMetadata"
• not sure why "cbvmda": "urn:epcglobal:cbv:mda" (which does not appear any most of those files) should be causing these files to fail validation...

[VladimirAlexiev]

all instances of "someFurtherMetadata" already appeared to be "ext1:someFurtherMetadata"

This is https://github.com/gs1/EPCIS/issues/276 issue4

not sure why "cbvmda": "urn:epcglobal:cbv:mda" (which does not appear any most of those files) should be causing these files to fail validation...

Because cbvmda is included in the context and the context says @protectd: true, so no other context can redefine that term, even if it defines it to the same namespace.

Since EPCIS JSON doesn't carry MasterData, I think we should remove this prefix from the context. @CraigRe @dakbhavesh @mgh128 do you agree? The ilms field is supposed to carry gs1: Voc properties, not cbvmda: properties, right?

[CraigRe]

@VladimirAlexiev up to now, EPCIS 1.x ILMD fields carry cbvmda properties. Even if we were allow EPCIS 2.0 ILMD fields to carry gs1: properties, we'll still need to support/allow cbvmda properties, to ensure backward compatibility with EPCIS 1.x implementations (and EPCIS Documents captured in those deployments).

[VladimirAlexiev]

I've seen cbvmda describing GLNs and the like (using cryptic names like ssa, sst). I now grepped "ilmd" and found several that use example:bestBeforeDate, and just one with "cbvmda:lotNumber": "LOTABC"

@CraigRe

• Can you point to more examples carrying cbvmda ILMD? Let's include them in the suite.
• Is cbvmda defined at some website like gs1 Voc is? I'll include such link in the READMEs

"cbvmda" does not appear any most of those files

Yes, but it appeared in a few XML that had some example master data. You can close this issue, since your comment shows the XML fix is correct.

[CraigRe]

@VladimirAlexiev those two examples (bestBeforeDate, lotNumber) are most common, in the DSCSA (Pharma) context. There is no online / linked data definition of cbvma.

[VladimirAlexiev]

• @sboeckelmann @dakbhavesh fixes needed: gs1:MT-Humidity is not valid anymore, should be gs1:Humidity or just Humidity (not sure which one). Appears in the following files (shown with count):

  grep -c '"gs1:' */*.xml|grep -v :0
  AssociationEvent/AssociationEventExamples.xml:1
  Mimasu/Example-associatonEvent-sensor.xml:1
  WithEventHashID/event_with_identical_hash_id_1.xml:1
  WithEventHashID/event_with_identical_hash_id_2.xml:1
  WithEventHashID/event_with_identical_hash_id_3.xml:1
  WithEventHashID/event_with_identical_hash_id_4.xml:1
  WithEventHashID/event_with_identical_hash_id_5.xml:1
  WithEventHashID/event_with_identical_hash_id_6.xml:1
  WithExtension/TransactionEvent.xml:1
  WithExtension/TransformationEvent.xml:1
  WithFullCombinationOfFields/aggregation_event_all_possible_fields.xml:1
  WithFullCombinationOfFields/association_event_all_possible_fields.xml:1
  WithFullCombinationOfFields/object_event_all_possible_fields.xml:1
  WithFullCombinationOfFields/transaction_event_all_possible_fields.xml:1
  WithFullCombinationOfFields/transformation_event_all_possible_fields.xml:1
  WithSensorData/SensorDataExamples.xml:46

• @CraigRe @RalphTro @sboeckelmann @dakbhavesh @mgh128 With the current definition, cbvmda:lotNumber is expanded to urn:epcglobal:cbv:mdalotNumber, and "mdalotNumber" is not a word. So I propose to add trailing colon to the namespace: urn:epcglobal:cbv:mda:. Fixes needed:
  • in JSONLD context
  • in XML:

    grep -c 'cbvmda' *.xml */*.xml|grep -v :0
    Example-PersistentDisposition.xml:1
    WithExtension/TransformationEvent.xml:2
    WithFullCombinationOfFields/object_event_all_possible_fields.xml:8
    WithFullCombinationOfFields/transformation_event_all_possible_fields.xml:8
    WithSensorData/SensorDataExamples.xml:2

• if the EPCIS context includes cbvmda then the XML->JSONLD convertor must skip that prefix (#276 issue 17)

@CraigRe please reopen and let's vote the last two bullets today

[VladimirAlexiev]

Decided in Sep 21 telco:

• gs1: values (Temperature, ERROR_CONDITION) should use full URLs in XML because no XML value uses CURIEs (they use full URNs or full URLs)
• The convertor should implement contraction in XML->JSON (down to the level of barewords), and expansion in JSON->XML
• We're keeping cbvmda: in the context
• The convertor should treat prefix cbvmda: just like epcis: cbv: gs1: omit them on XML->JSON, and add them on JSON->XML
• Craig and Elizabeth: GS1 US RxSecureSupplyChain WG uses cbvmda:itemExpirationDate and cbvmda:lotNumber as 2 of the key fields to meet DSCSA requirement. Can you confirm that in these discussions that we will NOT remove them? (confirmed)
• That WG uses other namespaces with no trailing delimiter, eg xmlns:gs1ushc="http://epcis.gs1us.org/hc/ns" and it's recommended to add one (in this case slash)

Discussion: whether to add to the cbvmda namespace a trailing colon, or trailing hash, or nothing

• Current HealthCare XMLs use hash for MasterData, and nothing for ILMD
• CBV 2.0 sec 9.3.1: "use hash for all props, but use colon for props site, sst, ssa, ssd for backward compatibility with CBV 1.2"
• Elizabeth and Neil will check with the HealthCare XML team
• Recommendation: use trailing slash for web URL namespaces, colon for URN namespaces (and don't use hash)
  • Vladimir: I added colon in my context: urn:epcglobal:cbv:mda:
  • @jmcanterafonseca-iota please do in your context when we have a decision
• @dakbhavesh please add a tweak in the XML 1.2->2.0 convertor to normalize to urn:epcglobal:cbv:mda:

[CraigRe]

@CraigRe reflect above in CBV § 9.3.1, using : instead of #.