language is extracted from @xx of rdfs:label

[tfrancart]

dataset.py maps the @XX of rdfs:label to the dct:language property:

            key = self.get_key(requested_key='rdfs:label')
            position = data.index(key) + 1
            description = data[position]

            languages = list()
            try:
                languages = [x[1].replace("@", "").lower() for x in description]

I don't think this is correct : the language of the content of a dataset might be different from the language(s) in which its labels are given. Instead a dct:language metadata should be sought.

[flopezag]

In the current version of the example files used for testing, this information is not presented, therefore we decided to select the corresponding language based on the language of the labels. In fact, languages make no reference to the language of the content but the languages of the different descriptions or texts provided in the entity.