Merge pull request #16 from TaskarCenterAtUW/feature-performance-improvement

Performance Improvement

Author: sujata-m

CHANGELOG.md

@@ -1,5 +1,9 @@
 # Change log
 
+### 0.2.3
+- Performance improvement if there are any errors
+
+
 ### 0.2.2
 - Added functionality to get the specific number of errors
   ```

src/example.py

@@ -5,6 +5,7 @@
 ASSETS_DIR = os.path.join(PARENT_DIR, 'tests/assets')
 VALID_ZIP_FILE = os.path.join(ASSETS_DIR, 'valid.zip')
 INVALID_ZIP_FILE = os.path.join(ASSETS_DIR, 'invalid.zip')
+INVALID_VANCOUVER_ZIP_FILE = os.path.join(ASSETS_DIR, 'vancouver-dataset.zip')
 SCHEMA_DIR = os.path.join(PARENT_DIR, 'src/python_osw_validation/schema')
 SCHEMA_FILE_PATH = os.path.join(SCHEMA_DIR, 'opensidewalks.schema.json')
 
@@ -24,15 +25,21 @@ def valid_test_without_provided_schema():
 def invalid_test_with_provided_schema():
     validator = OSWValidation(zipfile_path=INVALID_ZIP_FILE, schema_file_path=SCHEMA_FILE_PATH)
     result = validator.validate()
-    if not result.is_valid:
-        for error in result.errors:
-            print(error)
+    print(f'Number of errors: {len(result.errors)}')
     print(f'Invalid Test With Provided Schema: {"Failed" if result.is_valid else "Passed"}')
 
 
 def invalid_test_without_provided_schema():
     validator = OSWValidation(zipfile_path=INVALID_ZIP_FILE)
-    result = validator.validate()
+    result = validator.validate(max_errors=10)
+    print(f'Number of errors: {len(result.errors)}')
+    print(f'Invalid Test With Provided Schema: {"Failed" if result.is_valid else "Passed"}')
+
+
+def invalid_test_vancouver_dataset():
+    validator = OSWValidation(zipfile_path=INVALID_VANCOUVER_ZIP_FILE)
+    result = validator.validate(max_errors=30)
+    print(f'Number of errors: {len(result.errors)}')
     print(f'Invalid Test Without Schema: {"Failed" if result.is_valid else "Passed"}')
 
 
@@ -41,3 +48,4 @@ def invalid_test_without_provided_schema():
     valid_test_without_provided_schema()
     invalid_test_with_provided_schema()
     invalid_test_without_provided_schema()
+    invalid_test_vancouver_dataset()

src/python_osw_validation/__init__.py

@@ -152,13 +152,13 @@ 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 index, error in enumerate(errors):
-                if index < max_errors:
-                    self.errors.append(f'Validation error: {error.message}')
-                    if len(self.errors) == max_errors:
-                        break
+
+        for error in validator.iter_errors(geojson_data):
+            self.errors.append(f'Validation error: {error.message}')
+            if len(self.errors) == max_errors:
+                break
+
+        if len(self.errors) >= max_errors:
             return False
+
         return True

src/python_osw_validation/version.py

@@ -1 +1 @@
-__version__ = '0.2.2'
+__version__ = '0.2.3'