Dataset Review Feedback

specification/ai/Azure.AI.Projects/common/models.tsp

@@ -311,6 +311,9 @@ model PendingUploadRequest {
   @doc("If PendingUploadId is not provided, a random GUID will be used.")
   pendingUploadId?: string;
 
+  @doc("Name of Azure blob storage connection to use for generating temporary SAS token")
+  connectionName?: string;
+
   @doc("TemporaryBlobReference is the only supported type.")
   pendingUploadType: PendingUploadType.temporaryBlobReference;
 }
@@ -323,6 +326,9 @@ model PendingUploadResponse {
   @doc("ID for this upload request.")
   pendingUploadId: string;
 
+  @doc("Version of dataset to be created if user did not specify version when initially creating upload")
+  datasetVersion?: string;
+
   @doc("TemporaryBlobReference is the only supported type")
   pendingUploadType: PendingUploadType.temporaryBlobReference;
 }

specification/ai/Azure.AI.Projects/datasets/models.tsp

@@ -22,23 +22,46 @@ union DatasetType {
 
   @doc("URI file.")
   uri_file: "uri_file",
+
+  @doc("URI folder.")
+  uri_folder: "uri_folder",
 }
 
 #suppress "@azure-tools/typespec-azure-core/composition-over-inheritance"
 @doc("DatasetVersion Definition")
+@discriminator("type")
 model DatasetVersion {
   @doc("[Required] Uri of the data. Example: https://go.microsoft.com/fwlink/?linkid=2202330")
   @visibility("read", "create")
   @minLength(1)
   @pattern("[a-zA-Z0-9_]")
   datasetUri: string;
 
-  @doc("Data type")
-  datasetType: DatasetType;
+  @doc("Dataset type")
+  type: DatasetType;
+
+  @doc("Indicates if dataset is reference only or managed by dataset service. If true, the underlying data will be deleted when the dataset version is deleted")
+  @visibility("read")
+  isReference?: boolean;
 
   ...AssetBase;
 }
 
+@doc("FileDatasetVersion Definition")
+model FileDatasetVersion extends DatasetVersion {
+  @doc("Dataset type")
+  type: DatasetType.uri_file;
+
+  @doc("Indicates OpenAI Purpose. FileDatasets created with this field will be compatible with OpenAI-specific features")
+  openAIPurpose: string;
+}
+
+@doc("FileDatasetVersion Definition")
+model FolderDatasetVersion extends DatasetVersion {
+  @doc("Dataset type")
+  type: DatasetType.uri_folder;
+}
+
 @doc("PagedDataVersionBase Definition")
 @pagedResult
 model PagedDatasetVersion {

specification/ai/data-plane/Azure.AI.Projects/preview/2025-05-01-preview/azure-ai-projects-1dp.json

@@ -9480,7 +9480,8 @@
       "type": "string",
       "description": "Enum to determine the type of data.",
       "enum": [
-        "uri_file"
+        "uri_file",
+        "uri_folder"
       ],
       "x-ms-enum": {
         "name": "DatasetType",
@@ -9490,6 +9491,11 @@
             "name": "uri_file",
             "value": "uri_file",
             "description": "URI file."
+          },
+          {
+            "name": "uri_folder",
+            "value": "uri_folder",
+            "description": "URI folder."
           }
         ]
       }
@@ -9508,9 +9514,14 @@
             "create"
           ]
         },
-        "datasetType": {
+        "type": {
           "$ref": "#/definitions/DatasetType",
-          "description": "Data type"
+          "description": "Dataset type"
+        },
+        "isReference": {
+          "type": "boolean",
+          "description": "Indicates if dataset is reference only or managed by dataset service. If true, the underlying data will be deleted when the dataset version is deleted",
+          "readOnly": true
         },
         "stage": {
           "type": "string",
@@ -9553,9 +9564,10 @@
           "readOnly": true
         }
       },
+      "discriminator": "type",
       "required": [
         "datasetUri",
-        "datasetType"
+        "type"
       ]
     },
     "Deployment": {
@@ -9746,6 +9758,35 @@
         ]
       }
     },
+    "FileDatasetVersion": {
+      "type": "object",
+      "description": "FileDatasetVersion Definition",
+      "properties": {
+        "openAIPurpose": {
+          "type": "string",
+          "description": "Indicates OpenAI Purpose. FileDatasets created with this field will be compatible with OpenAI-specific features"
+        }
+      },
+      "required": [
+        "openAIPurpose"
+      ],
+      "allOf": [
+        {
+          "$ref": "#/definitions/DatasetVersion"
+        }
+      ],
+      "x-ms-discriminator-value": "uri_file"
+    },
+    "FolderDatasetVersion": {
+      "type": "object",
+      "description": "FileDatasetVersion Definition",
+      "allOf": [
+        {
+          "$ref": "#/definitions/DatasetVersion"
+        }
+      ],
+      "x-ms-discriminator-value": "uri_folder"
+    },
     "IdAssetReference": {
       "type": "object",
       "description": "Reference to an asset via its ARM resource ID.",
@@ -10134,6 +10175,10 @@
           "type": "string",
           "description": "If PendingUploadId is not provided, a random GUID will be used."
         },
+        "connectionName": {
+          "type": "string",
+          "description": "Name of Azure blob storage connection to use for generating temporary SAS token"
+        },
         "pendingUploadType": {
           "type": "string",
           "description": "TemporaryBlobReference is the only supported type.",
@@ -10161,6 +10206,10 @@
           "type": "string",
           "description": "ID for this upload request."
         },
+        "datasetVersion": {
+          "type": "string",
+          "description": "Version of dataset to be created if user did not specify version when initially creating upload"
+        },
         "pendingUploadType": {
           "type": "string",
           "description": "TemporaryBlobReference is the only supported type",