-Add external extensions to ExtractedDataValidator

-Validate external extensions against basic Open Geospatial Consortium (OGC) standards

Author: wisamuw

src/python_osw_validation/__init__.py

@@ -159,6 +159,17 @@ def validate(self) -> ValidationResult:
                     self.errors.append(f"Invalid {osw_file} geometries found, id's of invalid geometries: {set(invalid_geojson['_id'])}")
                     return ValidationResult(False, self.errors)
 
+            # Validate OSW external extensions
+            for file in validator.externalExtensions:
+                file_path = os.path.join(file)
+                extensionFile = gpd.read_file(file_path)
+                invalid_geojson = extensionFile[extensionFile.is_valid==False]
+                is_valid = len(invalid_geojson) == 0
+                if not is_valid:
+                    zip_handler.remove_extracted_files()
+                    self.errors.append(f"Invalid geometries found in extension file {file}, list of invalid geometries: {invalid_geojson.to_json()}")
+                    return ValidationResult(False, self.errors)
+            
             return ValidationResult(True)
         except Exception as e:
             self.errors.append(f'Unable to validate: {e}')

src/python_osw_validation/extracted_data_validator.py

@@ -6,6 +6,7 @@ class ExtractedDataValidator:
     def __init__(self, extracted_dir: str):
         self.extracted_dir = extracted_dir
         self.files = []
+        self.externalExtensions = []
         self.error = None
 
     def is_valid(self) -> bool:
@@ -43,5 +44,8 @@ def is_valid(self) -> bool:
         if required_files:
             self.error = f'Missing required .geojson files: {", ".join(required_files)}.'
             return False
+        
+        # Add OSW external extensions, GeoJSON files we know nothing about
+        self.externalExtensions.extend([item for item in geojson_files if item not in self.files])
 
         return True