Moved duplicated_property_types to utils and added tests

bam_masterdata/cli/cli.py

@@ -8,14 +8,14 @@
 from openpyxl import Workbook
 from rdflib import Graph
 
-from bam_masterdata.cli.duplicated_property_types import duplicated_property_types
 from bam_masterdata.cli.entities_to_excel import entities_to_excel
 from bam_masterdata.cli.entities_to_json import entities_to_json
 from bam_masterdata.cli.entities_to_rdf import entities_to_rdf
 from bam_masterdata.cli.fill_masterdata import MasterdataCodeGenerator
 from bam_masterdata.logger import logger
 from bam_masterdata.utils import (
     delete_and_create_dir,
+    duplicated_property_types,
     import_module,
     listdir_py_modules,
 )

bam_masterdata/utils/__init__.py

@@ -1,6 +1,7 @@
 from .utils import (
     code_to_class_name,
     delete_and_create_dir,
+    duplicated_property_types,
     import_module,
     listdir_py_modules,
     load_validation_rules,

bam_masterdata/utils/utils.py

@@ -1,7 +1,9 @@
 import glob
 import importlib.util
+import inspect
 import json
 import os
+import re
 import shutil
 from itertools import chain
 from typing import TYPE_CHECKING, Any, Optional
@@ -160,3 +162,44 @@ def load_validation_rules(
     except json.JSONDecodeError as e:
         logger.error(f"Error parsing validation rules JSON: {e}")
         raise ValueError(f"Error parsing validation rules JSON: {e}")
+
+
+from pathlib import Path
+
+
+def duplicated_property_types(module_path: str, logger: "BoundLoggerLazyProxy") -> dict:
+    """
+    Find the duplicated property types in a module specified by `module_path` and returns a dictionary
+    containing the duplicated property types class names as keys and the lines where they matched as values.
+
+    Args:
+        module_path (str): The path to the module containing the property types.
+        logger (BoundLoggerLazyProxy): The logger to log messages.
+
+    Returns:
+        dict: A dictionary containing the duplicated property types class names as keys and the
+        lines where they matched as values.
+    """
+    duplicated_props: dict = {}
+    module = import_module(module_path=module_path)
+    source_code = inspect.getsource(module)
+    for name, _ in inspect.getmembers(module):
+        if name.startswith("_") or name == "PropertyTypeDef":
+            continue
+
+        pattern = rf"^\s*{name} *= *PropertyTypeDef"
+
+        # Find all matching line numbers
+        matches = [
+            i + 1  # Convert to 1-based index
+            for i, line in enumerate(source_code.splitlines())
+            if re.match(pattern, line)
+        ]
+        if len(matches) > 1:
+            duplicated_props[name] = matches
+    if duplicated_props:
+        logger.critical(
+            f"Found {len(duplicated_props)} duplicated property types. These are stored in a dictionary "
+            f"where the keys are the names of the variables in property_types.py and the values are the lines in the module: {duplicated_props}"
+        )
+    return duplicated_props

tests/data/utils/example_prop_types_1.py

@@ -0,0 +1,23 @@
+from bam_masterdata.metadata.definitions import PropertyTypeDef
+
+PropA = PropertyTypeDef(
+    code="PROPA",
+    description="""repeated property""",
+    data_type="VARCHAR",
+    property_label="A1",
+)
+
+
+PropB = PropertyTypeDef(
+    code="PROPB",
+    description="""non-repeated property""",
+    data_type="VARCHAR",
+    property_label="B",
+)
+
+PropA = PropertyTypeDef(
+    code="PROPA",
+    description="""repeated property""",
+    data_type="VARCHAR",
+    property_label="A2",
+)

tests/data/utils/example_prop_types_2.py

@@ -0,0 +1,16 @@
+from bam_masterdata.metadata.definitions import PropertyTypeDef
+
+PropA = PropertyTypeDef(
+    code="PROPA",
+    description="""non-repeated property""",
+    data_type="VARCHAR",
+    property_label="A",
+)
+
+
+PropB = PropertyTypeDef(
+    code="PROPB",
+    description="""non-repeated property""",
+    data_type="VARCHAR",
+    property_label="B",
+)

tests/utils/test_utils.py

@@ -10,6 +10,7 @@
 from bam_masterdata.utils import (
     code_to_class_name,
     delete_and_create_dir,
+    duplicated_property_types,
     import_module,
     listdir_py_modules,
     load_validation_rules,
@@ -62,7 +63,12 @@ def test_delete_and_create_dir(
             "warning",
         ),
         # No Python files found in the directory
-        ("./tests/data", [], "No Python files found in the directory.", "info"),
+        (
+            "./tests/data/empty",
+            [],
+            "No Python files found in the directory.",
+            "info",
+        ),
         # Python files found in the directory
         (
             "./tests/utils",
@@ -270,3 +276,19 @@ def test_load_validation_rules(
         assert result == expected_output
         assert cleared_log_storage[-1]["event"] == expected_log
         assert cleared_log_storage[-1]["level"] == "info"
+
+
+@pytest.mark.parametrize(
+    "path, result",
+    [
+        # PropA appears twice
+        ("tests/data/utils/example_prop_types_1.py", {"PropA": [3, 18]}),
+        # None duplicated
+        ("tests/data/utils/example_prop_types_2.py", {}),
+    ],
+)
+def test_duplicated_property_types(cleared_log_storage: list, path: str, result: dict):
+    assert result == duplicated_property_types(path, logger)
+    if result:
+        assert cleared_log_storage[0]["level"] == "critical"
+        assert "Found 1 duplicated property types" in cleared_log_storage[0]["event"]