Array format not validating unicode markers in json

[IdrissaD]

I try to validate a csv file against a table schema, and I realized that unicode indicators u were breaking the validation of my arrays.

Schema:

 {
         "name":"communes",
         "description":"Liste des communes traversées par la randonnée",
         "example":"[{'nom': 'Molezon', 'code_INSEE': '48098'}]",
         "type":"array",
         "items":{
         	"type":"object",
         	"properties":{
         		"nom": {
         			"type":"string"
         		},
         		"code_INSEE":{
         			"type":"integer"
         		}
         	}
         }
      }

Value of communes field in my file: [{u'nom': u'Ispagnac', u'code_INSEE': u'48075'}]

Error:
type-error Type error in the cell "[{u'nom': u'Ispagnac', u'code_INSEE': u'48075'}]" in row "2" and field "communes" at position "9": type is "array/default"

Link to the workflow that show the errors, at lines 15, 16 and 17 (don't pay attention to the other errors, I fixed it): https://github.com/IdrissaD/schema_randonnee/runs/2381935747?check_suite_focus=true#step:6:15

Is it normal, and is there a way to make this work (apart from removing the u with a script) ?

(Edit: thanks for moving it here, didn't see the discussion tab at first...)

[roll]

@IdrissaD
Hi, I think during the data population process json.dumps has to be used because currently it's no a valid JSON because of u (looks like native Python stringification)

[IdrissaD]

Thanks for the tip! Yes it's Python stringification, I don't know why but I thought it was a valid JSON.