Telco 22. OGC Best Practice

When: WED 9 February 2022 - 15:00-16:00 CET

Where: GoToMeeting https://global.gotomeeting.com/join/321982654

Attendants

• OGC Members meeting:
  • 122nd Meeting: 28 February - 4 March 2022 (Virtual)
    • Citizen Science DWG session on 28 February at 21:45 CET: OGC Best Practice for using SensorThings API with Citizen Science (presentation and discussion) by Andreas Matheus, Joan Masó
    • SensorThings SWG session TBA
• Progress report on STAplus implementation
  • OGC Best Practices report submission

• Recap of Ownership in STA+ data model
  • Personal entities: Party, Thing, Datastream, Observation, Group can be deleted/updated by the linked Party.
  • Global entities: the rest can be created by any Party but cannot be deleted/updated.
• Andreas presented ACTION items for the OGC Best Practices document.
  • CREAF (BP 7.1): Creating (semantically enriched) data from Natusfera – how to demonstrate the interoperability and that you can get “the same” information from STAplus that you get from Natusfera
  • SECD (BP 7.2): STAplus latest and greatest achievements (Face Detection with Obervation Upload, Size and max Frames detection), Running a Camera Trap that uploads to STAplus
  • Fraunhofer (BP 7.3): Leveraging the STA API batch-processing
  • INRIA (BP7.4): Performance “issues” and the use of additional indexing
  • CREAF (BP7.5): Building a Web-Application to visualize STAplus
  • 52N (BP7.6): Modelling community processes (blogging)
  • SECD/INRIA/CREAF (future work): demonstrating the use of Microsoft Excel with OData/STAplus (e.g. Pl@ntNet or Natusfera data to the SECD endpoint)
    • INRIA will try loading millions of Pla@ntNet records from OData/STAplus to evaluate query performances in Excel - also to check any limit to data capacity in Excel
• DISCUSSION: batch-processing initialization and runtime
  • Observation (identification) should link the URI (e.g. GBIF taxonomy) to FeatureOfInterest, using Relation?
    • The role of Relation can be dwc:identificationReferences (http://rs.tdwg.org/dwc/terms/identificationReferences)
      • e.g. Subject = Observation (taxonID5219243), externalObject = FeatureOfInterest (https://www.gbif.org/species/5219243)
  • Can a user change the Datastream-> License?
    • Is it allowed to change a license on a datastream that already has observations?
  • What does it mean if the user sets the Group-> Runtime property?
    • A Group is “open” until the Runtime property defines a period [<start>, <end> = null]
      • “open”: Observations (and Relations) can be added to the Group by any Party
      • “closed”: Modifications to the Group can only be made by the owning Party (each update resets the <endTime>)
  • Does the Group need a closed property (License) to seal a set of observations and how strict should it be?
    • Among multiple parties, when one party deletes the individual account, its observations may stay in Group (as a read-only copy) but the owner information is N/A (anonymous)?
    • When an observation record is added to the Group, the service provider should warn or provide options about making a copy of the record?
      • This may be different from one service provider to another. In any case, how data update/delete is being handled should be made clear to the users.
    • Within the Group (sealed by License), observations shall be accessed (read-only) and used for training a dataset but cannot be updated/deleted?
      • e.g. Pl@ntNet AI service GBIF-DL: how was the license issue handled for loading training images from GBIF database?
    • Overview of the license hierarchy and compatibility

• AOB
  • STA+ data model will remain as it is in the UML diagram. Use cases fit to the STA+ will be presented as BPs.
  • Proposed submission by the end of May towards the 123rd OGC Member Meeting in mid-June 2022

• Pending discussion topics to be continued (at later dates):
  • Joan wanted to discuss the convenience to extend API model to include DQ_element in Datastream or in Group. currently it is only in observation, limiting the applicability of data quality to individual observations but and not to overall summaries.
  • Andreas was interested in discussing if observations can be edited and by whom.
    • There are obvious cases where it is done in practice when the series is recalibrated. Is this a new observation or an edit?
    • Need to record When, who and why the observation changed. Blockchain technologies are good at recoding this.
  • Kaori discussed on mapping Darwin Core and STA.
    • Where and how the taxon should be recorded?
      • Alternative A: Taxonomies could be considered observations and they could be related to the actual observations
      • Alternative B: The observed taxon could be encoded as a URI result that points to the full description of the taxon
      • Alternative C: "Relation" should allow for linking to external resources and taxon could be one of them. It seems the preferred alternative but this will require editing the STAplus data model.

• Progress on implementations of STAplus services and clients (Cos4Cloud)
  • Henning Bredel to demonstrate the 52North demo server: https://github.com/52North/sensorweb-server-sta/tree/feature/sta-plus/dao-citsci/docs
  • Andreas (SecureDimensions) has implemented the data cloud deployed by EGI, EOSC service provider: https://cos4cloud.secd.eu/staplus/v1.1/api (for Cos4Cloud project)
  • Joan (CREAF) has written the script to test exporting data from the Natusfera to the STAplus: https://github.com/joanma747/Natusfera4STAplus
  • Inria, one of the Cos4Cloud partners, has tested performance levels by exporting data (7 million records) from their Pl@ntNet to the STAplus, and will be testing to improve indexing in the database with Andreas: DB Performance | FROST-Server Documentation (fraunhoferiosb.github.io)
  • Semantic interoperability (e.g. RDF from W3C ) between those different data models (e.g. Natusfera, Pl@ntNet) applied to the STAplus model using the entity 'Relation'. An RDF query language such as SPARQL can be used then. Relation may play a critical role to explain taxonomy used in the biodiversity domain model. (e.g. Darwin Core terms, GBIF IPT User Manual Occurrence Data)
    • Kaori (CREAF) tested Relation using Darwin Core terms (e.g. dwc: recordedBy, dwc: identifiedBy, dwc: inCollection, dwc: toTaxon) in the STAplus.
    • Andreas suggested using OGC Definition Server (defs.opengis.net/vocprez/) to express Relation as well, which would strengthen our invention.
• Progress on the Definition Server: OGC platform to store vocabrary (linked to 'definition' in ObservedProperty): OGC Vocabularies (opengis.net)

• To discuss and explore the potential of exporting metadata (e.g. STAC, ISO, W3C DECAT) from various platforms into the STAplus once the common data model (template) is mapped.
  • e.g. GBIF Metadata Template for biodiversity domain
• Other activities existing and harmonization: https://core.citizenscience.org/ will be done later.

Connections

• New channel on Gitter:
• Usual channel in GoToMeeting number:
• Or use our email interop-cop3@listsNOspamPlease.weobserve.eu

GoToMeeting details

https://global.gotomeeting.com/join/321982654


You can also dial in using your phone.
(For supported devices, tap a one-touch number below to join instantly.)

United States: +1 (909) 259-0010
- One-touch: tel:+19092590010,,321982654#

Access Code: 321-982-654


More phone numbers:
(For supported devices, tap a one-touch number below to join instantly.)

Australia: +61 2 8014 4937
- One-touch: tel:+61280144937,,321982654#

Austria: +43 7 2088 1406
- One-touch: tel:+43720881406,,321982654#

Belgium: +32 28 08 4296
- One-touch: tel:+3228084296,,321982654#

Denmark: +45 69 91 89 33
- One-touch: tel:+4569918933,,321982654#

Finland: +358 942 59 9121
- One-touch: tel:+358942599121,,321982654#

France: +33 182 880 462
- One-touch: tel:+33182880462,,321982654#

Germany: +49 898 7806 6465
- One-touch: tel:+4989878066465,,321982654#

Ireland: +353 14 845 982
- One-touch: tel:+35314845982,,321982654#

Italy: +39 0 247 92 12 41
- One-touch: tel:+390247921241,,321982654#

Netherlands: +31 208 080 384
- One-touch: tel:+31208080384,,321982654#

Norway: +47 21 04 29 12
- One-touch: tel:+4721042912,,321982654#

Spain: +34 911 82 9783
- One-touch: tel:+34911829783,,321982654#

Sweden: +46 852 503 473
- One-touch: tel:+46852503473,,321982654#

Switzerland: +41 225 3314 54
- One-touch: tel:+41225331454,,321982654#

United Kingdom: +44 20 3318 4724
- One-touch: tel:+442033184724,,321982654#

1.1leveraging the STA API Batch-Processing