[Python] Inconsistent python generation of union creator types

[jbland64]

Naming convention for defining the union creator function

    const auto enum_fn = namer_.Function(enum_def);
    code += "\n";
    code += "def " + enum_fn + "Creator(unionType, table):";

https://github.com/google/flatbuffers/blob/master/src/idl_gen_python.cpp#L2594

Differs to the naming convention for calling it

    auto union_type = namer_.Type(enum_def);

    if (parser_.opts.include_dependence_headers) {
      union_type = namer_.NamespacedType(enum_def) + "." + union_type;
    }
    code += GenIndents(2) + "self." + field_field + " = " + union_type +
            "Creator(" + "self." + field_field + "Type, " + struct_var + "." +
            field_method + "())";
  }

https://github.com/google/flatbuffers/blob/master/src/idl_gen_python.cpp#L2029

I solved this locally by changing:

auto union_type = namer_.Type(enum_def);
to:
auto union_type = namer_.Function(enum_def);

I have not accounted for

    if (parser_.opts.include_dependence_headers) {
      union_type = namer_.NamespacedType(enum_def) + "." + union_type;
    }

This may need an additional method

  std::string NamespacedFunction(const Definition& def) const {
    return NamespacedString(def.defined_namespace, Function(def.name));
  }

but i have not tested this the name namespace functionality

[jtdavis777]

looks like a similar/same problem as #8791

[jbland64]

looks like a similar/same problem as #8791

Yes its the same problem just applied to a different data type, union vs struct

[jtdavis777]

@jbland64 would you be up for opening a PR to fix this? Possibly adding a test or two for it?

[jbland64]

Sure will do on Monday