[PLT-63] Add additional AWS validation (#1422)

Author: mnoszczak

labelbox/schema/dataset.py

@@ -153,6 +153,11 @@ def convert_field_keys(items):
                 "DataRow.row_data missing when creating DataRow.")
 
         row_data = args[DataRow.row_data.name]
+
+        if isinstance(row_data, str) and row_data.startswith("s3:/"):
+            raise InvalidQueryError(
+                "row_data: s3 assets must start with 'https'.")
+
         if not isinstance(row_data, str):
             # If the row data is an object, upload as a string
             args[DataRow.row_data.name] = json.dumps(row_data)
@@ -425,6 +430,10 @@ def validate_keys(item):
                 raise InvalidQueryError(
                     "`row_data` missing when creating DataRow.")
 
+            if isinstance(item.get('row_data'),
+                          str) and item.get('row_data').startswith("s3:/"):
+                raise InvalidQueryError(
+                    "row_data: s3 assets must start with 'https'.")
             invalid_keys = set(item) - {
                 *{f.name for f in DataRow.fields()}, 'attachments', 'media_type'
             }

tests/integration/test_data_rows.py

@@ -127,6 +127,18 @@ def test_get_data_row(data_row, client):
     assert client.get_data_row(data_row.uid)
 
 
+def test_create_invalid_aws_data_row(dataset, client):
+    with pytest.raises(labelbox.exceptions.InvalidQueryError) as exc:
+        dataset.create_data_row(row_data="s3://labelbox-public-data/invalid")
+    assert "s3" in exc.value.message
+
+    with pytest.raises(labelbox.exceptions.InvalidQueryError) as exc:
+        dataset.create_data_rows([{
+            "row_data": "s3://labelbox-public-data/invalid"
+        }])
+    assert "s3" in exc.value.message
+
+
 def test_lookup_data_rows(client, dataset):
     uid = str(uuid.uuid4())
     # 1 external id : 1 uid