OpenSearch re-indexing failed

[FuhuXia]

A lot of datasets failed to in indexed today.

OUT Re-indexing 2415 updated datasets...
OUT Batch 1/3: indexing 1000 dataset(s)...
OUT Warning: 1000 dataset(s) failed to index in this batch.
OUT {'dataset_id': '4a9db702-ecbe-4464-8771-a6ddf87b1560', 'status_code': 400, 'error_type':
'mapper_parsing_exception', 'error_reason': "failed to parse field [dcat.spatial] of type [text]
in document with id '4a9db702-ecbe-4464-8771-a6ddf87b1560'. Preview of field's value:
'{@type=Location, prefLabel=United States}'", 'caused_by': {'type': 'illegal_state_exception',
'reason': "Can't get text on a START_OBJECT at 1:790"}}

[FuhuXia]

All error datasets belong to source ssa-json. The reason is that they have a newly populated spatial field with value

"spatial": [{"@type": "Location", "prefLabel": "United States"}]

[FuhuXia]

Another reason is distribution.conformsTo value changed from

https://project-open-data.cio.gov/v1.1/schema

to

[{"@type": "Standard", "identifier": "https://www.w3.org/TR/tabular-data-primer/", "title": "CSV on the Web"}]

[jbrown-xentity]

@FuhuXia might be time to handle this holistically. Looks 3.0 to me.
I would by default make sure if the schema allows for an object but we don't want an object in opensearch to stringify the object by default.

[FuhuXia]

@jbrown-xentity
The PR is updated with a nested _normalize_dcat_metadata_value function to walk and fix any object/list values

[FuhuXia]

Fix is deployed on prod.