Type mismatch from Julia

[hhoeschle]

Hey,

we call the spinedb_api from Julia in the following way:

@show id_list = [Int16(db_map.get_or_add_object_class(name=obj_class)[1]) for obj_class in object_classes]
@show typeof(id_list)
@show relationship_class = db_map.add_wide_relationship_class(name = relationship_name, object_class_id_list=id_list)

Unfortunately, when submitting the id_list to python it is translated in a numpy.int64 rather than an int. This leads to the problem that the relationship is not correctly created, as the object_class_id in the sql-db is stored as BLOB.

Would it be possible to integrate the following explicit typecast:

            for wide_kwargs in wide_kwargs_list:
                for dimension, object_class_id in enumerate(wide_kwargs["object_class_id_list"]):
                    narrow_kwargs = {
                        "id": id,
                        "dimension": dimension,
                        "object_class_id": int(object_class_id),
                        "name": wide_kwargs["name"],
                    }
                    item_list.append(narrow_kwargs)

I guess it may also occur at other places. I will collect and report them below.

Thanks
Hanspeter

Link to file

[hhoeschle]

Also

"object_id": int(object_id)

Link to file

[manuelma]

perhaps we should cast all ids to int...

[manuelma]

Have you checked if PyCall says anything about this ?

[hhoeschle]

The casting is done here

I guess somehow PyLong_FromLongLong maps onto numpy.int64 ... but it is hard to conclude

[jkiviluo]

Must be an outdated issue.