Check names of additional attribute keys (#196)

* Add check for illegal additional attribute keys

* Add test with illegal additional attribute name

* Add two more examples of illegal attribute keys

* Fix validator after merging main

* Fix test after merging main

nomenclature/code.py

@@ -1,7 +1,8 @@
 import re
+from keyword import iskeyword
 from typing import Any, Dict, List, Optional, Set, Union
 
-from pydantic import BaseModel, Field
+from pydantic import BaseModel, Field, validator
 
 
 class Code(BaseModel):
@@ -11,6 +12,18 @@ class Code(BaseModel):
     description: Optional[str]
     extra_attributes: Dict[str, Any] = {}
 
+    @validator("extra_attributes")
+    def check_attribute_names(cls, v, values):
+        # Check that attributes only contains keys which are valid identifiers
+        if illegal_keys := [
+            key for key in v.keys() if not key.isidentifier() or iskeyword(key)
+        ]:
+            raise ValueError(
+                "Only valid identifiers are allowed as attribute keys. Found "
+                f"'{illegal_keys}' in '{values['name']}' which are not allowed."
+            )
+        return v
+
     @classmethod
     def from_dict(cls, mapping) -> "Code":
         if isinstance(mapping, str):

tests/test_code.py

@@ -1,6 +1,6 @@
 import pytest
 
-from nomenclature.code import VariableCode
+from nomenclature.code import Code, VariableCode
 
 
 def test_variable_without_unit_raises():
@@ -23,6 +23,13 @@ def test_variable_alias_setting():
     )
 
 
+@pytest.mark.parametrize("illegal_key", ["contains-hyphen", "also not allowed", "True"])
+def test_illegal_additional_attribute(illegal_key):
+    match = f"{illegal_key}.*'code1'.*not allowed"
+    with pytest.raises(ValueError, match=match):
+        Code(name="code1", extra_attributes={illegal_key: True})
+
+
 def test_variable_multiple_units():
     """Test that a VariableCode with multiple units works"""
     var = VariableCode.from_dict({"Var1": {"unit": ["unit1", "unit2"]}})