add default value type checking

.travis.yml

@@ -65,6 +65,7 @@ matrix:
     - language: python
       python: "3.6"
       env: TOXENV=py36
+      install: pip3 install jsonschema==3.0.1
       script:
         # DSL tests
         - cd $TRAVIS_BUILD_DIR/sdk/python

samples/notebooks/DSL Static Type Checking.ipynb

@@ -195,7 +195,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 1,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -255,7 +255,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 2,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -393,7 +393,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 5,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -417,7 +417,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": 6,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -477,7 +477,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": 7,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -507,7 +507,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 9,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -565,7 +565,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
+   "execution_count": 10,
    "metadata": {},
    "outputs": [
     {
@@ -600,7 +600,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 12,
+   "execution_count": 11,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -630,7 +630,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 12,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -687,7 +687,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 14,
+   "execution_count": 13,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -710,7 +710,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 15,
+   "execution_count": 14,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -733,9 +733,18 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 15,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Integer has a property openapi_schema_validator that the latter does not.\n",
+      "Component \"task_factory_a\" is expecting field_o to be type(Integer), but the passed argument is type({'Integer': {'openapi_schema_validator': {'type': 'integer'}}})\n"
+     ]
+    }
+   ],
    "source": [
     "@component\n",
     "def task_factory_a(field_m: {'GCSPath': {'openapi_schema_validator': '{\"type\": \"string\", \"pattern\": \"^gs://.*$\"}'}}, field_o: 'Integer'):\n",
@@ -751,7 +760,7 @@
     "# Pipeline input types are also checked against the component I/O types.\n",
     "@dsl.pipeline(name='type_check_g',\n",
     "    description='')\n",
-    "def pipeline_g(a: {'GCSPath': {'openapi_schema_validator': '{\"type\": \"string\", \"pattern\": \"^gs://.*$\"}'}}='good', b: Integer()=12):\n",
+    "def pipeline_g(a: {'GCSPath': {'openapi_schema_validator': '{\"type\": \"string\", \"pattern\": \"^gs://.*$\"}'}}='gs://kfp-path', b: Integer()=12):\n",
     "    task_factory_a(field_m=a, field_o=b)\n",
     "\n",
     "try:\n",
@@ -769,7 +778,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 17,
+   "execution_count": 16,
    "metadata": {},
    "outputs": [],
    "source": [

samples/notebooks/Lightweight Python components - basics.ipynb

@@ -45,7 +45,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 1,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -61,7 +61,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 2,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -80,7 +80,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 3,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -96,7 +96,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 4,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -154,9 +154,20 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 9,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "MyDivmodOutput(quotient=14, remainder=2)"
+      ]
+     },
+     "execution_count": 9,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
    "source": [
     "my_divmod(100, 7)"
    ]
@@ -172,7 +183,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 6,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -189,7 +200,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 7,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -223,7 +234,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 8,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -278,7 +289,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.5.4rc1"
+   "version": "3.6.5"
   }
  },
  "nbformat": 4,

sdk/python/kfp/dsl/_metadata.py

@@ -70,7 +70,10 @@ def serialize(self):
 
   @staticmethod
   def deserialize(payload):
-    # If the payload is a string of a dict serialization, convert it back to a dict
+    '''deserialize expects two types of input: dict and str
+    1) If the payload is a string of a dict serialization, convert it back to a dict
+    2) If the payload is a string, the type is named as such with no properties.
+    3) If the payload is a dict, the type name and properties are extracted. '''
     try:
       import ast
       payload = ast.literal_eval(payload)
@@ -138,6 +141,9 @@ def _annotation_to_typemeta(annotation):
   '''_annotation_to_type_meta converts an annotation to an instance of TypeMeta
   Args:
     annotation(BaseType/str/dict): input/output annotations
+      BaseType: registered in kfp.dsl.types
+      str: either a string of a dict serialization or a string of the type name
+      dict: type name and properties. note that the properties values can be dict.
   Returns:
     TypeMeta
     '''
@@ -221,6 +227,14 @@ def _extract_pipeline_metadata(func):
     arg_default = arg_defaults[arg] if arg in arg_defaults else None
     if arg in annotations:
       arg_type = _annotation_to_typemeta(annotations[arg])
+    if 'openapi_schema_validator' in arg_type.properties and arg_default is not None:
+      from jsonschema import validate
+      import json
+      schema_object = arg_type.properties['openapi_schema_validator']
+      if isinstance(schema_object, str):
+        # In case the property value for the schema validator is a string instead of a dict.
+        schema_object = json.loads(arg_type.properties['openapi_schema_validator'])
+      validate(instance=arg_default, schema=schema_object)
     pipeline_meta.inputs.append(ParameterMeta(name=arg, description='', param_type=arg_type, default=arg_default))
 
   #TODO: add descriptions to the metadata

sdk/python/kfp/dsl/types.py

@@ -19,65 +19,76 @@ class BaseType:
 
 # Primitive Types
 class Integer(BaseType):
-	openapi_schema_validator = {
-		"type": "integer"
-	}
+	def __init__(self):
+		self.openapi_schema_validator = {
+			"type": "integer"
+		}
 
 class String(BaseType):
-	openapi_schema_validator = {
-		"type": "string"
-	}
+	def __init__(self):
+		self.openapi_schema_validator = {
+			"type": "string"
+		}
 
 class Float(BaseType):
-	openapi_schema_validator = {
-		"type": "number"
-	}
+	def __init__(self):
+		self.openapi_schema_validator = {
+			"type": "number"
+		}
 
 class Bool(BaseType):
-	openapi_schema_validator = {
-		"type": "boolean"
-	}
+	def __init__(self):
+		self.openapi_schema_validator = {
+			"type": "boolean"
+		}
 
 class List(BaseType):
-	openapi_schema_validator = {
-		"type": "array"
-	}
+	def __init__(self):
+		self.openapi_schema_validator = {
+			"type": "array"
+		}
 
 class Dict(BaseType):
-	openapi_schema_validator = {
-		"type": "object",
-	}
+	def __init__(self):
+		self.openapi_schema_validator = {
+			"type": "object",
+		}
 
 # GCP Types
 class GCSPath(BaseType):
-	openapi_schema_validator = {
-		"type": "string",
-		"pattern": "^gs://.*$"
-	}
+	def __init__(self):
+		self.openapi_schema_validator = {
+			"type": "string",
+			"pattern": "^gs://.*$"
+		}
 
 class GCRPath(BaseType):
-	openapi_schema_validator = {
-		"type": "string",
-		"pattern": "^.*gcr\\.io/.*$"
-	}
+	def __init__(self):
+		self.openapi_schema_validator = {
+			"type": "string",
+			"pattern": "^.*gcr\\.io/.*$"
+		}
 
 class GCPRegion(BaseType):
-	openapi_schema_validator = {
-		"type": "string"
-	}
+	def __init__(self):
+		self.openapi_schema_validator = {
+			"type": "string"
+		}
 
 class GCPProjectID(BaseType):
 	'''MetaGCPProjectID: GCP project id'''
-	openapi_schema_validator = {
-		"type": "string"
-	}
+	def __init__(self):
+		self.openapi_schema_validator = {
+			"type": "string"
+		}
 
 # General Types
 class LocalPath(BaseType):
 	#TODO: add restriction to path
-	openapi_schema_validator = {
-		"type": "string"
-	}
+	def __init__(self):
+		self.openapi_schema_validator = {
+			"type": "string"
+		}
 
 class InconsistentTypeException(Exception):
 	'''InconsistencyTypeException is raised when two types are not consistent'''

sdk/python/requirements.txt

@@ -11,3 +11,4 @@ google-auth>=1.6.1
 requests_toolbelt>=0.8.0
 kfp-server-api >= 0.1.18, < 0.1.19
 argo-models == 2.2.1a
+jsonschema >= 3.0.1