Update ComponentGraph to enforce needing .x and .y for each component specified in the graph

docs/source/release_notes.rst

@@ -13,6 +13,7 @@ Release Notes
         * Moved ``get_hyperparameter_ranges`` to ``PipelineBase`` class from automl/utils module :pr:`2546`
         * Renamed ``ComponentGraph``'s ``get_parents`` to ``get_inputs`` :pr:`2540`
         * Removed ``ComponentGraph.linearized_component_graph`` and ``ComponentGraph.from_list`` :pr:`2556`
+        * Updated ``ComponentGraph`` to enforce requiring `.x` and `.y` inputs for each component in the graph :pr:`2563`
     * Documentation Changes
         * Improved detail of ``TextFeaturizer`` docstring and tutorial :pr:`2568`
     * Testing Changes
@@ -24,6 +25,7 @@ Release Notes
         * Moved ``get_hyperparameter_ranges`` to ``PipelineBase`` class from automl/utils module :pr:`2546`
         * Renamed ``ComponentGraph``'s ``get_parents`` to ``get_inputs`` :pr:`2540`
         * Removed ``ComponentGraph.linearized_component_graph`` and ``ComponentGraph.from_list`` :pr:`2556`
+        * Updated ``ComponentGraph`` to enforce requiring `.x` and `.y` inputs for each component in the graph :pr:`2563`
 
 
 **v0.29.0 Jul. 21, 2021**

docs/source/user_guide/pipelines.ipynb

@@ -52,11 +52,11 @@
    "outputs": [],
    "source": [
     "component_graph_as_dict = {\n",
-    "    'Imputer': ['Imputer'],\n",
-    "    'Encoder': ['One Hot Encoder', 'Imputer'],\n",
-    "    'Random Forest Clf': ['Random Forest Classifier', 'Encoder'],\n",
-    "    'Elastic Net Clf': ['Elastic Net Classifier', 'Encoder'],\n",
-    "    'Final Estimator': ['Logistic Regression Classifier', 'Random Forest Clf', 'Elastic Net Clf']\n",
+    "    'Imputer': ['Imputer', 'X', 'y'],\n",
+    "    'Encoder': ['One Hot Encoder', 'Imputer.x', 'y'],\n",
+    "    'Random Forest Clf': ['Random Forest Classifier', 'Encoder.x', 'y'],\n",
+    "    'Elastic Net Clf': ['Elastic Net Classifier', 'Encoder.x', 'y'],\n",
+    "    'Final Estimator': ['Logistic Regression Classifier', 'Random Forest Clf.x', 'Elastic Net Clf.x', 'y']\n",
     "}\n",
     "\n",
     "MulticlassClassificationPipeline(component_graph=component_graph_as_dict)"
@@ -205,11 +205,11 @@
    "outputs": [],
    "source": [
     "component_graph_as_dict = {\n",
-    "    'Imputer': ['Imputer'],\n",
-    "    'Encoder': ['One Hot Encoder', 'Imputer'],\n",
-    "    'Random Forest Clf': ['Random Forest Classifier', 'Encoder'],\n",
-    "    'Elastic Net Clf': ['Elastic Net Classifier', 'Encoder'],\n",
-    "    'Final Estimator': ['Logistic Regression Classifier', 'Random Forest Clf', 'Elastic Net Clf']\n",
+    "    'Imputer': ['Imputer', 'X', 'y'],\n",
+    "    'Encoder': ['One Hot Encoder', 'Imputer.x', 'y'],\n",
+    "    'Random Forest Clf': ['Random Forest Classifier', 'Encoder.x', 'y'],\n",
+    "    'Elastic Net Clf': ['Elastic Net Classifier', 'Encoder.x', 'y'],\n",
+    "    'Final Estimator': ['Logistic Regression Classifier', 'Random Forest Clf.x', 'Elastic Net Clf.x', 'y']\n",
     "}\n",
     "\n",
     "nonlinear_pipeline = MulticlassClassificationPipeline(component_graph=component_graph_as_dict)\n",
@@ -477,4 +477,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 4
-}
+}

evalml/pipelines/component_graph.py

@@ -39,20 +39,37 @@ def __init__(self, component_dict=None, random_seed=0):
             raise ValueError(
                 "component_dict must be a dictionary which specifies the components and edges between components"
             )
+        self._validate_component_dict()
         self.component_instances = {}
         self._is_instantiated = False
         for component_name, component_info in self.component_dict.items():
-            if not isinstance(component_info, list):
-                raise ValueError(
-                    "All component information should be passed in as a list"
-                )
             component_class = handle_component_class(component_info[0])
             self.component_instances[component_name] = component_class
         self.input_feature_names = {}
         self._feature_provenance = {}
         self._i = 0
         self._compute_order = self.generate_order(self.component_dict)
 
+    def _validate_component_dict(self):
+        for _, component_inputs in self.component_dict.items():
+            if not isinstance(component_inputs, list):
+                raise ValueError(
+                    "All component information should be passed in as a list"
+                )
+            component_inputs = component_inputs[1:]
+            has_feature_input = any(
+                component_input.endswith(".x") or component_input == "X"
+                for component_input in component_inputs
+            )
+            has_target_input = any(
+                component_input.endswith(".y") or component_input == "y"
+                for component_input in component_inputs
+            )
+            if not (has_feature_input and has_target_input):
+                raise ValueError(
+                    "All edges must be specified as either an input feature (.x) or input target (.y)."
+                )
+
     @property
     def compute_order(self):
         """The order that components will be computed or called in."""
@@ -280,7 +297,7 @@ def _compute_features(self, component_list, X, y=None, fit=False):
                     output = component_instance.predict(input_x)
                 else:
                     output = None
-                output_cache[component_name] = output
+                output_cache[f"{component_name}.x"] = output
         return output_cache
 
     def _get_feature_provenance(self, input_feature_names):
@@ -522,6 +539,7 @@ def generate_order(cls, component_dict):
         if len(edges) == 0:
             return []
         digraph = nx.DiGraph()
+        digraph.add_nodes_from(list(component_dict.keys()))
         digraph.add_edges_from(edges)
         if not nx.is_weakly_connected(digraph):
             raise ValueError("The given graph is not completely connected")