Merge PR #15 Add array content in scoring output schema

from task/scoring-output-schema-array-content

Author: instanceofme

CHANGELOG.md

@@ -1,5 +1,8 @@
 # Changelog
 
+## Version 1.1.1 (2024-05-07)
+* Scoring recipe: Specify array content type for decision_rules in output schema
+
 ## Version 1.1.0 (2023-05-11)
 * Webapp: Fix autosplit (outdated kwarg in sklearn's DecisionTreeClassifier)
 * Evaluation recipe: Fix recipe with multiclass tasks + import error

plugin.json

@@ -1,6 +1,6 @@
 {
   "id" : "decision-tree-builder",
-  "version" : "1.1.0",
+  "version" : "1.1.1",
   "meta" : {
     "label" : "Interactive Decision Tree Builder",
     "description" : "Build and explore decision trees, and use them to score and evaluate data",

python-lib/dku_idtb_scoring/score.py

@@ -18,7 +18,10 @@ def update_input_schema(input_schema, columns):
     return new_input_schema
 
 def _add_column(name, type, schema, columns=None):
-    schema.append({'type': type, 'name': name})
+    col = {'type': type, 'name': name}
+    if type == 'array':
+        col['arrayContent'] = {'type': 'string'}
+    schema.append(col)
     if columns is not None:
         columns.append(name)
 

python-tests/test_score.py

@@ -122,7 +122,7 @@ def test_scored_df_schema():
 		{"type": "double", "name": "proba_A"},
 		{"type": "double", "name": "proba_B"},
 		{"type": "string", "name": "prediction"},
-		{"type": "array", "name": "decision_rule"},
+		{"type": "array", "name": "decision_rule", "arrayContent": {"type": "string"}},
 		{"type": "int", "name": "leaf_id"},
 		{"type": "string", "name": "label"}
 	]
@@ -131,7 +131,7 @@ def test_scored_df_schema():
 	schema = get_scored_df_schema(tree, get_input_schema(), columns, False, True, False)
 	expected_schema = [
 		{"type": "string", "name": "prediction"},
-		{"type": "array", "name": "decision_rule"},
+		{"type": "array", "name": "decision_rule", "arrayContent": {"type": "string"}},
 		{"type": "int", "name": "leaf_id"},
 		{"type": "string", "name": "label"}
 	]
@@ -144,7 +144,7 @@ def test_scored_df_schema():
 		{"type": "double", "name": "num"},
 		{"type": "string", "name": "prediction"},
 		{"type": "boolean", "name": "prediction_correct"},
-		{"type": "array", "name": "decision_rule"},
+		{"type": "array", "name": "decision_rule", "arrayContent": {"type": "string"}},
 		{"type": "int", "name": "leaf_id"},
 		{"type": "string", "name": "label"}
 	]