Fix for Task 1078

- Fix for Task [1078](https://dev.azure.com/TDEI-UW/TDEI/_workitems/edit/1078)
- Added the functionality to stop processing further if there are 20 errors(default)
- max_errors can be passed as a parameter to validate function
- Added unit test cases
- Updated README.md
- Updated CHANGELOG.md
- Updated Package version

Author: sujata-m

CHANGELOG.md

@@ -1,5 +1,13 @@
 # Change log
 
+### 0.2.2
+- Added functionality to get the specific umber of errors
+  ```
+  validator = OSWValidation(zipfile_path=<ZIP_FILE_PATH>)
+  result = validator.validate() // will return only first 20 errors by default
+  result = validator.validate(max_errors=10) // will return only first 10 errors
+  ```
+
 ### 0.2.1
 - Updated zipfile_handler
 - Fixed "No .geojson files found in the specified directory or its subdirectories." issue

README.md

@@ -28,9 +28,13 @@ This package validates the OSW geojson file. Package requires a OSW zip file pat
 from python_osw_validation import OSWValidation
 
 validator = OSWValidation(zipfile_path='<Zip file path>')
-result = validator.validate()
+result = validator.validate()  
 print(result.is_valid)
-print(result.errors)
+print(result.errors) # will return first 20 errors by default if there are errors
+
+result = validator.validate(max_errors=10)  
+print(result.is_valid)
+print(result.errors) # will return first 10 errors depending on the max_errors parameter
 
 ```
 

src/python_osw_validation/__init__.py

@@ -45,7 +45,7 @@ def are_ids_unique(self, gdf):
 
         return is_valid, list(duplicates)
 
-    def validate(self) -> ValidationResult:
+    def validate(self, max_errors=20) -> ValidationResult:
         try:
             # Extract the zipfile
             zip_handler = ZipFileHandler(self.zipfile_path)
@@ -62,7 +62,8 @@ def validate(self) -> ValidationResult:
                 return ValidationResult(False, self.errors)
             for file in validator.files:
                 file_path = os.path.join(file)
-                is_valid = self.validate_osw_errors(file_path)
+                if not self.validate_osw_errors(file_path, max_errors):
+                    break
 
             if self.errors:
                 zip_handler.remove_extracted_files()
@@ -72,7 +73,8 @@ def validate(self) -> ValidationResult:
             OSW_dataset = {}
             for file in validator.files:
                 file_path = os.path.join(file)
-                osw_file = next((osw_file_any for osw_file_any in OSW_dataset_files.keys() if osw_file_any in file_path), '')
+                osw_file = next(
+                    (osw_file_any for osw_file_any in OSW_dataset_files.keys() if osw_file_any in file_path), '')
                 OSW_dataset[osw_file] = gpd.read_file(file_path)
 
             # Are all id's unique in each file? No need to check uniqueness across files yet since we do not have a global OSW ID format yet
@@ -103,35 +105,42 @@ def validate(self) -> ValidationResult:
             unmatched = node_ids_edges_u - node_ids
             is_valid = len(unmatched) == 0
             if not is_valid:
-                self.errors.append(f"All _u_id's in edges should be part of _id's mentioned in nodes, _u_id's not in nodes are: {unmatched}")
+                self.errors.append(
+                    f"All _u_id's in edges should be part of _id's mentioned in nodes, _u_id's not in nodes are: {unmatched}")
 
             # Do all node references in _v_id exist in nodes?
             unmatched = node_ids_edges_v - node_ids
             is_valid = len(unmatched) == 0
             if not is_valid:
-                self.errors.append(f"All _v_id's in edges should be part of _id's mentioned in nodes, _v_id's not in nodes are: {unmatched}")
+                self.errors.append(
+                    f"All _v_id's in edges should be part of _id's mentioned in nodes, _v_id's not in nodes are: {unmatched}")
 
             # Do all node references in _w_id exist in nodes?
             unmatched = node_ids_zones_w - node_ids
             is_valid = len(unmatched) == 0
             if not is_valid:
-                self.errors.append(f"All _w_id's in zones should be part of _id's mentioned in nodes, _w_id's not in nodes are: {unmatched}")
+                self.errors.append(
+                    f"All _w_id's in zones should be part of _id's mentioned in nodes, _w_id's not in nodes are: {unmatched}")
 
             # Geometry validation: check geometry type in each file and test if coordinates make a shape that is reasonable geometric shape according to the Simple Feature Access standard
             for osw_file in OSW_dataset:
-                invalid_geojson = OSW_dataset[osw_file][(OSW_dataset[osw_file].geometry.type != OSW_dataset_files[osw_file]['geometry']) | (OSW_dataset[osw_file].is_valid==False)]
+                invalid_geojson = OSW_dataset[osw_file][
+                    (OSW_dataset[osw_file].geometry.type != OSW_dataset_files[osw_file]['geometry']) | (
+                          OSW_dataset[osw_file].is_valid == False)]
                 is_valid = len(invalid_geojson) == 0
                 if not is_valid:
-                    self.errors.append(f"Invalid {osw_file} geometries found, id's of invalid geometries: {set(invalid_geojson['_id'])}")
+                    self.errors.append(
+                        f"Invalid {osw_file} geometries found, id's of invalid geometries: {set(invalid_geojson['_id'])}")
 
             # 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]
+                invalid_geojson = extensionFile[extensionFile.is_valid == False]
                 is_valid = len(invalid_geojson) == 0
                 if not is_valid:
-                    self.errors.append(f"Invalid geometries found in extension file {file}, list of invalid geometries: {invalid_geojson.to_json()}")
+                    self.errors.append(
+                        f"Invalid geometries found in extension file {file}, list of invalid geometries: {invalid_geojson.to_json()}")
 
             if self.errors:
                 zip_handler.remove_extracted_files()
@@ -147,14 +156,17 @@ def load_osw_file(self, graph_geojson_path: str) -> Dict[str, Any]:
         with open(graph_geojson_path, 'r') as file:
             return json.load(file)
 
-    def validate_osw_errors(self, file_path: str) -> bool:
+    def validate_osw_errors(self, file_path: str, max_errors: int) -> bool:
         '''Validate OSW Data against the schema and process all errors'''
         geojson_data = self.load_osw_file(file_path)
         validator = jsonschema.Draft7Validator(self.load_osw_schema(self.schema_file_path))
         errors = list(validator.iter_errors(geojson_data))
 
         if errors:
-            for error in errors:
-                self.errors.append(f'Validation error: {error.message}')
+            for index, error in enumerate(errors):
+                if index < max_errors:
+                    self.errors.append(f'Validation error: {error.message}')
+                    if len(self.errors) == max_errors:
+                        break
             return False
         return True

src/python_osw_validation/version.py

@@ -1 +1 @@
-__version__ = '0.2.1'
+__version__ = '0.2.2'

tests/unit_tests/test_osw_validation.py

@@ -76,6 +76,20 @@ def test_invalid_zipfile_with_schema(self):
         self.assertFalse(result.is_valid)
         self.assertIsNotNone(result.errors)
 
+    def test_invalid_zipfile_default_error_count(self):
+        validation = OSWValidation(zipfile_path=self.invalid_zipfile)
+        result = validation.validate()
+        self.assertFalse(result.is_valid)
+        self.assertIsNotNone(result.errors)
+        self.assertEqual(20, len(result.errors))
+
+    def test_invalid_zipfile_should_return_10_errors(self):
+        validation = OSWValidation(zipfile_path=self.invalid_zipfile)
+        result = validation.validate(max_errors=10)
+        self.assertFalse(result.is_valid)
+        self.assertIsNotNone(result.errors)
+        self.assertEqual(10, len(result.errors))
+
     def test_invalid_zipfile_with_invalid_schema(self):
         validation = OSWValidation(zipfile_path=self.invalid_zipfile,
                                        schema_file_path=self.invalid_schema_file_path)