Attributes declarations

Could you review this proposal @bduye and @pdi ?

Into dataset.json if attributes_labels is specified, we should add an attributes_values_labels field (following the declaration of attributes_labels) as a new field covering all the values that we should find in the additional column of the observations.tsv: The attributes_labels declares the additional column header so the attributes_values_labels should declare the value contained in this additionnal column cells.

{
    "attributes_labels": {
        "OBS_STATUS": "Observation Status",
    },
    "attributes_values_labels":{
        "OBS_STATUS": {
            "F": "Fixed",
            "P": "Provisionnal",
            "E": "Estimated"
        }
    }

Moreover, we should declare into series.json the attributes following the dimensions if attributes are declared. But this should be discussed as attribute is set for one value of the timeseries as attributes could be empty and attributes is not mandatory:

• attributes may be list all the possibilities inside this columns such as:

"series": [
        {
            "code": "M.MIO_M3.AL.4100.AT",
            "dimensions": {
                "freq": "M",
                "geo": "AT",
                "partner": "AL",
                "product": "4100",
                "unit": "MIO_M3"
            },
            "attributes": {
                "OBSV_STATUS": [
                    "P",
                    "E"
                ]
            }
        },
        {
            "code": "M.MIO_M3.AL.4100.BE",
            "dimensions": {
                "freq": "M",
                "geo": "BE",
                "partner": "AL",
                "product": "4100",
                "unit": "MIO_M3"
            }
        }

or just the simple declaration of the attributes ? As if no value inside the cell => no attributes

"attributes": [
                "OBSV_STATUS",
            ]