Add schema links to error messages resolution section

sdk/ml/azure-ai-ml/azure/ai/ml/_exception_helper.py

@@ -154,32 +154,32 @@ def format_errors_and_resolutions_sections(entity_type: str, error_types: Dict[s
     if error_types[ValidationErrorType.INVALID_VALUE]:
         errors += f"\n{count}) One or more fields are invalid"
         resolutions += f"Double-check that all specified parameters are of the correct types and formats \
-        prescribed by the {entity_type} schema.\n"
+        prescribed by the {entity_type} schema."
         count += 1
     if error_types[ValidationErrorType.UNKNOWN_FIELD]:
         errors += f"\n{count}) A least one unrecognized parameter is specified"
-        resolutions += f"Remove any parameters not prescribed by the {entity_type} schema.\n"
+        resolutions += f"Remove any parameters not prescribed by the {entity_type} schema."
         count += 1
     if error_types[ValidationErrorType.MISSING_FIELD]:
         errors += f"\n{count}) At least one required parameter is missing"
-        resolutions += f"Ensure all parameters required by the {entity_type} schema are specified.\n"
+        resolutions += f"Ensure all parameters required by the {entity_type} schema are specified."
         count += 1
     if error_types[ValidationErrorType.FILE_OR_FOLDER_NOT_FOUND]:
         errors += f"\n{count}) One or more files or folders do not exist.\n"
-        resolutions += "Double-check the directory paths you provided and enter the correct paths.\n"
+        resolutions += "Double-check the directory paths you provided and enter the correct paths."
         count += 1
     if error_types[ValidationErrorType.CANNOT_SERIALIZE]:
         errors += f"\n{count}) One or more fields cannot be serialized.\n"
         resolutions += f"Double-check that all specified parameters are of the correct types and formats \
-        prescribed by the {entity_type} schema.\n"
+        prescribed by the {entity_type} schema."
         count += 1
     if error_types[ValidationErrorType.CANNOT_PARSE]:
         errors += f"\n{count}) YAML file cannot be parsed.\n"
-        resolutions += "Double-check your YAML file for syntax and formatting errors.\n"
+        resolutions += "Double-check your YAML file for syntax and formatting errors."
         count += 1
     if error_types[ValidationErrorType.RESOURCE_NOT_FOUND]:
         errors += f"\n{count}) Resource was not found.\n"
-        resolutions += "Double-check that the resource has been specified correctly and that you have access to it.\n"
+        resolutions += "Double-check that the resource has been specified correctly and that you have access to it."
         count += 1
 
     return errors, resolutions
@@ -191,10 +191,50 @@ def format_create_validation_error(
     """
     Formats a detailed error message for validation errors.
     """
+    from azure.ai.ml.entities._util import REF_DOC_ERROR_MESSAGE_MAP
+    from azure.ai.ml._schema.assets.data import DataSchema
+    from azure.ai.ml._schema._datastore import (
+        AzureBlobSchema,
+        AzureDataLakeGen1Schema,
+        AzureDataLakeGen2Schema,
+        AzureFileSchema,
+    )
+    from azure.ai.ml._schema.job import CommandJobSchema
+    from azure.ai.ml._schema._sweep import SweepJobSchema
+    from azure.ai.ml._schema.assets.environment import EnvironmentSchema
+    from azure.ai.ml._schema.assets.model import ModelSchema
+
     entity_type, details = get_entity_type(error)
     error_types, details = format_details_section(error, details, entity_type)
     errors, resolutions = format_errors_and_resolutions_sections(entity_type, error_types)
-    description = YAML_CREATION_ERROR_DESCRIPTION.format(entity_type=entity_type) if yaml_operation else ""
+
+    if yaml_operation:
+        description = YAML_CREATION_ERROR_DESCRIPTION.format(entity_type=entity_type)
+
+        if entity_type == ErrorTarget.MODEL:
+            schema_type = ModelSchema
+        elif entity_type == ErrorTarget.DATA:
+            schema_type = DataSchema
+        elif entity_type == ErrorTarget.COMMAND_JOB:
+            schema_type = CommandJobSchema
+        elif entity_type == ErrorTarget.SWEEP_JOB:
+            schema_type = SweepJobSchema
+        elif entity_type in [ErrorTarget.BLOB_DATASTORE, ErrorTarget.DATASTORE]:
+            schema_type = AzureBlobSchema
+        elif entity_type == ErrorTarget.GEN1_DATASTORE:
+            schema_type = AzureDataLakeGen1Schema
+        elif entity_type == ErrorTarget.GEN2_DATASTORE:
+            schema_type = AzureDataLakeGen2Schema
+        elif entity_type == ErrorTarget.FILE_DATASTORE:
+            schema_type = AzureFileSchema
+        elif entity_type == ErrorTarget.ENVIRONMENT:
+            schema_type = EnvironmentSchema
+
+        resolutions += " " + REF_DOC_ERROR_MESSAGE_MAP.get(schema_type, "")
+    else:
+        description = ""
+
+    resolutions += "\n"
     formatted_error = SCHEMA_VALIDATION_ERROR_TEMPLATE.format(
         description=description,
         error_msg=errors,

sdk/ml/azure-ai-ml/azure/ai/ml/constants/_common.py

@@ -102,7 +102,7 @@
 EXPERIMENTAL_LINK_MESSAGE = (
     "and may change at any time. Please see https://aka.ms/azuremlexperimental for more information."
 )
-REF_DOC_YAML_SCHEMA_ERROR_MSG_FORMAT = "For a more detailed breakdown of the {} schema, please see: {}."
+REF_DOC_YAML_SCHEMA_ERROR_MSG_FORMAT = "Visit this link to refer to the {} schema if needed: {}."
 STORAGE_AUTH_MISMATCH_ERROR = "AuthorizationPermissionMismatch"
 SWEEP_JOB_BEST_CHILD_RUN_ID_PROPERTY_NAME = "best_child_run_id"
 BATCH_JOB_CHILD_RUN_NAME = "batchscoring"

sdk/ml/azure-ai-ml/azure/ai/ml/exceptions.py

@@ -64,6 +64,10 @@ class ErrorTarget:
     ONLINE_ENDPOINT = "OnlineEndpoint"
     ASSET = "Asset"
     DATASTORE = "Datastore"
+    BLOB_DATASTORE = "BlobDatastore"
+    FILE_DATASTORE = "FileDatastore"
+    GEN1_DATASTORE = "Gen1Datastore"
+    GEN2_DATASTORE = "Gen2Datastore"
     WORKSPACE = "Workspace"
     COMPUTE = "Compute"
     DEPLOYMENT = "Deployment"

sdk/ml/azure-ai-ml/tests/dataset/unittests/test_data_operations.py

@@ -13,7 +13,12 @@
     DatasetVersionDetails,
 )
 from azure.ai.ml._scope_dependent_operations import OperationConfig, OperationScope
-from azure.ai.ml.constants._common import AssetTypes
+from azure.ai.ml.constants._common import (
+    REF_DOC_YAML_SCHEMA_ERROR_MSG_FORMAT,
+    AssetTypes,
+    YAMLRefDocLinks,
+    YAMLRefDocSchemaNames,
+)
 from azure.ai.ml.entities._assets import Data
 from azure.ai.ml.entities._assets._artifacts.artifact import ArtifactStorageInfo
 from azure.ai.ml.exceptions import ErrorTarget
@@ -170,13 +175,17 @@ def test_create_or_update_missing_path(self, mock_data_operations: DataOperation
         Expect to raise ValidationException for missing path
         """
         name = "random_name"
-        data = Data(name=name, version="1", description="this is an mltable dataset", type=AssetTypes.MLTABLE)
+        data1 = Data(name=name, version="1", description="this is an mltable dataset", type=AssetTypes.MLTABLE)
 
         with pytest.raises(Exception) as ex:
-            mock_data_operations.create_or_update(data)
+            mock_data_operations.create_or_update(data1)
         assert "At least one required parameter is missing" in str(ex.value)
         mock_data_operations._operation.create_or_update.assert_not_called()
 
+        with pytest.raises(Exception) as ex:
+            load_data("tests/test_configs/dataset/data_missing_path_test.yml")
+        assert REF_DOC_YAML_SCHEMA_ERROR_MSG_FORMAT.format(YAMLRefDocSchemaNames.DATA, YAMLRefDocLinks.DATA) in str(ex.value)
+
     @patch("azure.ai.ml.operations._data_operations.read_local_mltable_metadata_contents")
     @patch("azure.ai.ml.operations._data_operations.read_remote_mltable_metadata_contents")
     @patch("azure.ai.ml.operations._data_operations.download_mltable_metadata_schema")

sdk/ml/azure-ai-ml/tests/internal_utils/unittests/test_asset_utils.py

@@ -64,7 +64,11 @@ def link_file_path(
     link_file_name = "link_file_rand_name.txt"
     link_file = Path(os.path.join(os.path.abspath(storage_test_directory), link_file_name))
 
-    os.symlink(target_file_path, link_file)
+    try:
+        os.symlink(target_file_path, link_file)
+    except FileExistsError:
+        pass
+
     assert os.path.islink(link_file)
     link_file = convert_windows_path_to_unix(link_file)
     yield link_file