Split the first intervention into a casual observation and the intervention itself

[PietrH]

Request by @LienReyserhove and Karel:

---

The first intervention for an event can be identified:

For a Dossier_ID, sort on Laatst_Bewerkt_Datum, the first date where Dossier_status == Opvoling -> This is the first intervention. This is when the species was observed for the very first time for that event.

We want to split this record into 2 records:

1. If there is an action: this is an intervention and should be mapped to samplingProtocol as it is currently
2. Also, always, the records should be mapped to samplingProtocol == casual observation

So you can have only the casual observation, but you can't have the intervention without the casual observation.

Any further interventions, don't get a casual observation.

---

This is rather urgent.

[PietrH]

Discussed with @damianooldoni

It's probably better on the early alert side to only send emails for:

datasetkey + occurrenceID + eventID is new

However, eventID is not always present for other datasets.

So if eventID is present, only send emails for new combinations of datasetkey and eventID.

He will discuss with Nico

[PietrH]

We will need a new occurrenceID for the newly created occurrence record, this can be the original id + a suffix: -obs -cas or a GUID

It's not an ideal solution, but we can still try

[PietrH]

This will take a bit more work than I expected, as it's going to require restructuring the current mapping a little bit.

[PietrH]

When creating new occurrences from interventions, the casual observation occurrence if you will, what fields should these new records "inherit" from their parent record?

In other words, what columns should be retained when I create a new record based on the first intervention in an event?

---

@LienReyserhove Are there any columns that only make sense for an intervention occurrence record? That should be empty for an observation occurrence record?

[PietrH]

Discussed with @damianooldoni

It's probably better on the early alert side to only send emails for:

datasetkey + occurrenceID + eventID is new

However, eventID is not always present for other datasets.

So if eventID is present, only send emails for new combinations of datasetkey and eventID.

He will discuss with Nico

Discussed with @LienReyserhove :

• Observations should go Early Alert (first interventions, and observations)
• Interventions should go to MICA dashboard.

These can be told apart based on samplingProtocol == "casual observation"

@damianooldoni : to be discussed with Nico

[PietrH]

Currently no longer in the pipeline, waiting for new input from Lien/RATO to go ahead