Update examples

Author: ospillinger

examples/fraud/implementations/models/dnn.py

@@ -1,9 +1,9 @@
-def transform_spark(data, columns, args, transformed_column_name):
+def transform_spark(data, input, transformed_column_name):
     import pyspark.sql.functions as F
 
-    distribution = args["class_distribution"]
+    distribution = input["class_distribution"]
 
     return data.withColumn(
         transformed_column_name,
-        F.when(data[columns["col"]] == 0, distribution[1]).otherwise(distribution[0]),
+        F.when(data[input["col"]] == 0, distribution[1]).otherwise(distribution[0]),
     )

examples/fraud/implementations/transformers/weight.py

@@ -1,5 +1,5 @@
 - kind: api
   name: fraud
-  model_name: dnn
+  model: @dnn
   compute:
     replicas: 1

examples/fraud/resources/apis.yaml

@@ -0,0 +1,19 @@
+- kind: model
+  name: dnn
+  estimator: cortex.dnn_classifier
+  target_column: @class
+  input:
+    num_classes: 2
+    numeric_columns:
+      [@time_normalized, @v1_normalized, @v2_normalized, @v3_normalized, @v4_normalized,
+       @v5_normalized, @v6_normalized, @v7_normalized, @v8_normalized, @v9_normalized,
+       @v10_normalized, @v11_normalized, @v12_normalized, @v13_normalized, @v14_normalized,
+       @v15_normalized, @v16_normalized, @v17_normalized, @v18_normalized, @v19_normalized,
+       @v20_normalized, @v21_normalized, @v22_normalized, @v23_normalized, @v24_normalized,
+       @v25_normalized, @v26_normalized, @v27_normalized, @v28_normalized, @amount_normalized]
+  training_input:
+    weight_column: @weight_column
+  hparams:
+    hidden_units: [100, 100, 100]
+  training:
+    num_steps: 5000

examples/fraud/resources/dnn.yaml

@@ -1,3 +1,24 @@
+- kind: template
+  name: normalize
+  yaml: |
+    - kind: aggregate
+      name: {column}_mean
+      aggregator: cortex.mean
+      input: @{column}
+
+    - kind: aggregate
+      name: {column}_stddev
+      aggregator: cortex.stddev
+      input: @{column}
+
+    - kind: transformed_column
+      name: {column}_normalized
+      transformer: cortex.normalize
+      input:
+        col: @{column}
+        mean: @{column}_mean
+        stddev: @{column}_stddev
+
 - kind: embed
   template: normalize
   args:

examples/fraud/resources/models.yaml

@@ -5,7 +5,7 @@
     path: s3a://cortex-examples/fraud.csv
     csv_config:
       header: true
-    schema: [time, v1, v2, v3, v4, v5, v6, v7, v8, v9, v10, v11, v12, v13, v14, v15, v16, v17, v18, v19, v20, v21, v22, v23, v24, v25, v26, v27, v28, amount, class]
+    schema: [@time, @v1, @v2, @v3, @v4, @v5, @v6, @v7, @v8, @v9, @v10, @v11, @v12, @v13, @v14, @v15, @v16, @v17, @v18, @v19, @v20, @v21, @v22, @v23, @v24, @v25, @v26, @v27, @v28, @amount, @class]
 
 - kind: raw_column
   name: time

examples/fraud/resources/normalize.yaml

@@ -1,24 +1,11 @@
 - kind: aggregate
   name: class_distribution
   aggregator: cortex.class_distribution_int
-  inputs:
-    columns:
-      col: class
-
-- kind: transformer
-  name: weight
-  inputs:
-    columns:
-      col: INT_COLUMN
-    args:
-      class_distribution: {INT: FLOAT}
-  output_type: FLOAT_COLUMN
+  input: @class
 
 - kind: transformed_column
   name: weight_column
-  transformer: weight
-  inputs:
-    columns:
-      col: class
-    args:
-      class_distribution: class_distribution
+  transformer_path: implementations/transformers/weight.py
+  input:
+    col: @class
+    class_distribution: @class_distribution