gbif / vocabulary

A simple registry of controlled vocabularies used for terms found in GBIF mediated data.
Apache License 2.0
5 stars 1 forks source link

Create initial draft of the eventType vocabulary #107

Closed timrobertson100 closed 2 years ago

timrobertson100 commented 2 years ago

There will soon be a proposal to add eventType as a term in Darwin Core, to be used in the event core schema and beyond. This will surely need to accommodate a large variety of concepts but over time an upper taxonomy may be achievable to structure these.

In order to support the GBIF/ALA developments to bring SamplingEvent search and the ongoing work to Diversify the Data Model we're working through a series of exemplar datasets which require this vocabulary.

I suggest we draft what is necessary for these case studies as a start in the form of a simple CSV containing concept, parentConcept, definitionInEnglish. This will be sufficient to quickly bootstrap a vocabulary in the UAT Vocab Server enabling the integration into the pipleines.

timrobertson100 commented 2 years ago

CC @charvolant

timrobertson100 commented 2 years ago

Please see https://registry.gbif-uat.org/vocabulary/EventType/concepts

This was created using

touch /tmp/empty.csv
java -cp vocabulary-importer/target/vocabulary-importer-0.49-SNAPSHOT.jar org.gbif.vocabulary.importer.CliApp \
--vocabularyName "EventType" \
--vocabularyLabelEN "Event Type" \
--vocabularyDefinitionEN "An early draft of a vocabulary for DwC:eventType, developed to support the pilot work on the ALA and GBIF. See https://github.com/gbif/vocabulary/issues/107" \
--conceptsPath "/tmp/eventType.csv" \
--csvDelimiter ";" \
--apiUrl "https://api.gbif-uat.org/v1/" \
--apiUser "trobertson" \
--hiddenLabelsPath "/tmp/empty.csv" \
--apiPassword

and eventType.csv which was provided by @charvolant

CecSve commented 5 days ago

Should MediaSample be a concept? It seems more like a BOR than the other concepts and I cannot come up with an example of when an event would happen because of 'media', as I can imagine for bioblitz, expedition etc.

Since the issue is closed, maybe this discussion need to happen elsewhere?

tucotuco commented 5 days ago

As MediaSample, no, as that indicates an instance of Media, but an Event to capture a MediaSample (such as a camera tap photo, does seems reasonable, but I would cast that as an Event, such as MediaSampling, or MediaCapture.

CecSve commented 4 days ago

As MediaSample, no, as that indicates an instance of Media, but an Event to capture a MediaSample (such as a camera tap photo, does seems reasonable, but I would cast that as an Event, such as MediaSampling, or MediaCapture.

Wouldn't deployment make sense for camera traps? Or I could get confused about the filename deployments in a camtrap dp archive. But I was thinking:

├── Event │ ├── Deployment │ │ └── MediaSampling/MediaCapture │ │ └── Video │ │ └── Audio │ │ └── Photograph

This would cover various equipment methods that generate a media output used for species detection.

timrobertson100 commented 4 days ago

Wouldn't deployment make sense for camera traps?

Yes, and could either reflect the event of setting it/them up allowing configuration information to be attached, or the full duration of the camera(s) while in position ("we deployed cameras for 3 months..."), or both when used in a hierarchy.

MediaCapture sounds like a good name for the type of events that produce audio, video or photograph entities.

tucotuco commented 4 days ago

This example highlights something interesting to try to be aware of going forward. The hierarchy in the eventType vocabulary must be a hierarchy of eventTypes, not a hierarchy of Events.

So, for example, a MediaCapture Event might have a Deployment parent Event, but that does not mean that a MediaCapture Event is a subtype of Deployment Event. I would consult the Camera Trap people for their understanding of this, but to me the MediaCapture is not a Deployment subtype.