Deploy Identify REST service

[cmungall]

We need to deploy an Identify REST service.

Input: lat, long, [optional] time; optional precision? + layer IDs Output: json payload for each requested layer

See also

• 66 - these mappings will be composed with REST payloads OR the REST service will use these to return ENVO for qualitative payloads

• 55

• 21 - once we have the service

Note: this ticket to be assigned to @usethedata or Yaxing (I invited Bruce to org, don't have GH for Yaxing)

[cmungall]

We can discuss here requirements and preferences for the service

My own preferences:

• should be described using openapi3/swagger (alternative: HATEOS)
• all IDs in the system should follow NMDC identifier policy (see #2)
• payloads should be JSON (ideally json-ld)

I'm imagining two main endpoints/routes:

• /layers/
  • input: none
  • output: returns list of layers. ideally each is identified by a unique ID, plus any necessary metadata about that layer (e.g {id: DAAC:1, name: "soil", description: ".....", ...}
• /identify/
  • Input:
    • lat, long,
      • [optional] time
      • [list] layer IDs
      • [optional] precision?
      • [optional] vocabulary e.g. ENVO [TBD: does the API map or does the client map?]
  • Output: json(ld)
    • one object per layer
    • layer ID
    • precision
    • EITHER
    • vocabulary name (e.g. USDA soil layer)
    • vocabulary term ID (e.g 12)
    • vocabular term value (e.g. "alluvial soil")
    • OR
    • unit + value
    • plus any other provenance info

Happy to advise on any implementation aspects (esp if Python/flask). @kfagnan aim3 may want to weigh in here

[wdduncan]

Start with the soil ZOBLER layer.
Global soil type in the tsv file/layer.

@Blancohl If possible add provence metadata about the layers. Who made it? When was it made?

[wdduncan]

Need to set up workflow for adding terms to ENVO.