Schema and dataset fixes based on import findings

[lydiascarf]

The action items as I understand them are to:

• update the RDLS v0.2 and v0.3 schemas to use anyOf instead of oneOf for date validation
• update the RDLS v0.3 schema to nest $ref under items for processes (see here)
• add occurrences to all datasets without occurrences on events
• drop the custom validation logic and run the standard jsonschema validator against all of our files
• align editor validations
  • don't include empty doi values in PR
  • check the spelling on collaborator role
• update hazard process mapping & re-import all datasets

See #48 (comment)

[matamadio]

I updated v0.3. v0.2 won't be used after update - all old datasets will be updated under v1.0

[bennyistanto]

I just checked the India GLOBS and ESA CCI Global Land cover, which already exist in the JKAN, and it appears the RDL exposure information is missing.

India GOBS

ESA CCI Land cover

Compared to the JRC LUISA dataset

@matamadio Does this problem due to I put "exp_item" per #339 which not yet implemented? Or something else?

@lydiascarf I also noticed that when I access https://jkan.riskdatalibrary.org/, it goes to the Dataset page instead of the Main page.

[matamadio]

@bennyistanto yes we removed the catalogue frontpage.
Also yes, the schema has not changed to your latest proposal.
Please make those datasets compliant to current state of schema, e.g.:

      "exposure": [
        {
          "metrics": [
            {
              "dimension": "structure",
              "quantity_kind": "monetary",
              "currency": "EUR",
              "id": "metric_uz6EGP14"
            },
            {
              "dimension": "content",
              "quantity_kind": "area",
              "currency": "",
              "id": "metric_vbOHCKjI"
            }
          ],
          "category": "buildings",
          "taxonomy": "GED4ALL",
          "id": "item_leHYkYah"
        }
      ]

Note I just pushed an update on the MDE fixing a few bugs on exposure.

[bennyistanto]

I understand your concern, but for some case, how to differentiate two exposure variable in single dataset without exp_item?

Example: coal plant and gas pipeline, where it hase same metric dimension: Structure, but it has different quantity kind: count and length

[matamadio]

I see your point; see new proposal #339

[matamadio]

Somehow the hazard information is not translated in this dataset: https://jkan.riskdatalibrary.org/datasets/rdls_hzd_aqueduct/

Yet the information is in the json - at the event_sets level.

https://github.com/GFDRR/rdl-jkan/blob/rdl-0.3/_datasets/json/rdls_hzd-wri_glb_water.json

      "hazard": {
        "event_sets": [
          {
            "hazards": [
              {
                "type": "drought",
                "hazard_process": "socioeconomic_drought",
                "id": "hazard_3rjfHwN9"
              }
            ],
            "analysis_type": "deterministic",
            "calculation_method": "simulated",
            "id": "event_set_63sWwbBH"
          }

[lydiascarf]

Somehow the hazard information is not translated in this dataset: https://jkan.riskdatalibrary.org/datasets/rdls_hzd_aqueduct/

Yet the information is in the json - at the event_sets level.

https://github.com/GFDRR/rdl-jkan/blob/rdl-0.3/_datasets/json/rdls_hzd-wri_glb_water.json

      "hazard": {
        "event_sets": [
          {
            "hazards": [
              {
                "type": "drought",
                "hazard_process": "socioeconomic_drought",
                "id": "hazard_3rjfHwN9"
              }
            ],
            "analysis_type": "deterministic",
            "calculation_method": "simulated",
            "id": "event_set_63sWwbBH"
          }

Fixed: https://jkan.riskdatalibrary.org/datasets/rdls_hzd-wri_glb_water/