refactor: decision metric to target metric

Darinochka · Darinochka · commit 0bcef43b7a7b · 2025-01-28T18:51:18.000+03:00
diff --git a/autointent/_datafiles/default-multiclass-config.yaml b/autointent/_datafiles/default-multiclass-config.yaml
@@ -1,14 +1,14 @@
 # TODO: make up a better and more versatile config
 - node_type: embedding
-  decision_metric: retrieval_hit_rate
+  target_metric: retrieval_hit_rate
   search_space:
     - module_name: retrieval
       k: [10]
       embedder_name:
         - avsolatorio/GIST-small-Embedding-v0
         - infgrad/stella-base-en-v2
 - node_type: scoring
-  decision_metric: scoring_roc_auc
+  target_metric: scoring_roc_auc
   search_space:
     - module_name: knn
       k: [1, 3, 5, 10]
@@ -20,7 +20,7 @@
         - cross-encoder/ms-marco-MiniLM-L-6-v2
       k: [1, 3, 5, 10]
 - node_type: decision
-  decision_metric: decision_accuracy
+  target_metric: decision_accuracy
   search_space:
     - module_name: threshold
       thresh: [0.5]
diff --git a/autointent/_datafiles/default-multilabel-config.yaml b/autointent/_datafiles/default-multilabel-config.yaml
@@ -1,20 +1,20 @@
 # TODO: make up a better and more versatile config
 - node_type: embedding
-  decision_metric: retrieval_hit_rate_intersecting
+  target_metric: retrieval_hit_rate_intersecting
   search_space:
     - module_name: retrieval
       k: [10]
       embedder_name:
         - deepvk/USER-bge-m3
 - node_type: scoring
-  decision_metric: scoring_roc_auc
+  target_metric: scoring_roc_auc
   search_space:
     - module_name: knn
       k: [3]
       weights: ["uniform", "distance", "closest"]
     - module_name: linear
 - node_type: decision
-  decision_metric: decision_accuracy
+  target_metric: decision_accuracy
   search_space:
     - module_name: threshold
       thresh: [0.5]
diff --git a/autointent/nodes/_optimization/_node_optimizer.py b/autointent/nodes/_optimization/_node_optimizer.py
@@ -38,7 +38,8 @@ def __init__(
         self.decision_metric_name = decision_metric
 
         self.metrics = metrics if metrics is not None else []
-        self.metrics.append(self.decision_metric_name)
+        if self.decision_metric_name not in self.metrics:
+            self.metrics.append(self.decision_metric_name)
 
         self.modules_search_spaces = search_space  # TODO search space validation
         self._logger = logging.getLogger(__name__)  # TODO solve duplicate logging messages problem
diff --git a/tests/assets/configs/description.yaml b/tests/assets/configs/description.yaml
@@ -1,16 +1,16 @@
 - node_type: embedding
-  decision_metric: retrieval_hit_rate
+  target_metric: retrieval_hit_rate
   search_space:
     - module_name: retrieval
       k: [10]
       embedder_name:
         - sentence-transformers/all-MiniLM-L6-v2
 - node_type: scoring
-  decision_metric: scoring_roc_auc
+  target_metric: scoring_roc_auc
   search_space:
     - module_name: description
       temperature: [1.0, 0.5, 0.1, 0.05]
 - node_type: decision
-  decision_metric: decision_accuracy
+  target_metric: decision_accuracy
   search_space:
     - module_name: argmax
diff --git a/tests/assets/configs/multiclass.yaml b/tests/assets/configs/multiclass.yaml
@@ -1,13 +1,13 @@
 - node_type: embedding
-  decision_metric: retrieval_hit_rate
+  target_metric: retrieval_hit_rate
   search_space:
     - module_name: retrieval
       k: [10]
       embedder_name:
         - sentence-transformers/all-MiniLM-L6-v2
         - avsolatorio/GIST-small-Embedding-v0
 - node_type: scoring
-  decision_metric: scoring_roc_auc
+  target_metric: scoring_roc_auc
   search_space:
     - module_name: knn
       k: [5, 10]
@@ -32,7 +32,7 @@
       cross_encoder_name:
         - cross-encoder/ms-marco-MiniLM-L-6-v2
 - node_type: decision
-  decision_metric: decision_accuracy
+  target_metric: decision_accuracy
   search_space:
     - module_name: threshold
       thresh: [0.5, [0.5, 0.5, 0.5, 0.5]]
diff --git a/tests/assets/configs/multilabel.yaml b/tests/assets/configs/multilabel.yaml
@@ -1,13 +1,13 @@
 - node_type: embedding
-  decision_metric: scoring_accuracy
+  target_metric: scoring_accuracy
   search_space:
     - module_name: logreg
       cv: [2]
       embedder_name:
         - sentence-transformers/all-MiniLM-L6-v2
         - avsolatorio/GIST-small-Embedding-v0
 - node_type: scoring
-  decision_metric: scoring_roc_auc
+  target_metric: scoring_roc_auc
   search_space:
     - module_name: knn
       k: [5, 10]
@@ -28,7 +28,7 @@
       cross_encoder_name:
         - cross-encoder/ms-marco-MiniLM-L-6-v2
 - node_type: decision
-  decision_metric: decision_accuracy
+  target_metric: decision_accuracy
   search_space:
     - module_name: threshold
       thresh: [0.5, [0.5, 0.5, 0.5, 0.5]]
diff --git a/tests/nodes/test_decision.py b/tests/nodes/test_decision.py
@@ -15,7 +15,7 @@ def test_decision_multiclass(scoring_optimizer_multiclass):
     scoring_optimizer_multiclass.fit(context)
 
     decision_optimizer_config = {
-        "metric": "decision_accuracy",
+        "target_metric": "decision_accuracy",
         "node_type": "decision",
         "search_space": [
             {"module_name": "threshold", "thresh": [0.5]},
@@ -54,7 +54,7 @@ def test_decision_multilabel(scoring_optimizer_multilabel):
     scoring_optimizer_multilabel.fit(context)
 
     decision_optimizer_config = {
-        "metric": "decision_accuracy",
+        "target_metric": "decision_accuracy",
         "node_type": "decision",
         "search_space": [
             {"module_name": "threshold", "thresh": [0.5]},
diff --git a/tests/nodes/test_logreg.py b/tests/nodes/test_logreg.py
@@ -56,7 +56,7 @@ def get_embedding_optimizer(multilabel: bool):
     if multilabel:
         metric = "scoring_neg_coverage"
     embedding_optimizer_config = {
-        "metric": metric,
+        "target_metric": metric,
         "node_type": "embedding",
         "search_space": [
             {
diff --git a/tests/nodes/test_retrieval.py b/tests/nodes/test_retrieval.py
@@ -54,7 +54,7 @@ def get_embedding_optimizer(multilabel: bool):
     if multilabel:
         metric = metric + "_intersecting"
     embedding_optimizer_config = {
-        "metric": metric,
+        "target_metric": metric,
         "node_type": "embedding",
         "search_space": [
             {
diff --git a/tests/nodes/test_scoring.py b/tests/nodes/test_scoring.py
@@ -16,7 +16,7 @@ def test_scoring_multiclass(embedding_optimizer_multiclass):
     embedding_optimizer_multiclass.fit(context)
 
     scoring_optimizer_config = {
-        "metric": "scoring_roc_auc",
+        "target_metric": "scoring_roc_auc",
         "node_type": "scoring",
         "search_space": [
             {
@@ -80,7 +80,7 @@ def test_scoring_multilabel(embedding_optimizer_multilabel):
     embedding_optimizer_multilabel.fit(context)
 
     scoring_optimizer_config = {
-        "metric": "scoring_roc_auc",
+        "target_metric": "scoring_roc_auc",
         "node_type": "scoring",
         "search_space": [
             {

Original file line number	Diff line number	Diff line change
`@@ -56,7 +56,7 @@ def get_embedding_optimizer(multilabel: bool):`
`56`	`56`	`if multilabel:`
`57`	`57`	`metric = "scoring_neg_coverage"`
`58`	`58`	`embedding_optimizer_config = {`
`59`		`- "metric": metric,`
	`59`	`+ "target_metric": metric,`
`60`	`60`	`"node_type": "embedding",`
`61`	`61`	`"search_space": [`
`62`	`62`	`{`
Original file line number	Diff line number	Diff line change
`@@ -54,7 +54,7 @@ def get_embedding_optimizer(multilabel: bool):`
`54`	`54`	`if multilabel:`
`55`	`55`	`metric = metric + "_intersecting"`
`56`	`56`	`embedding_optimizer_config = {`
`57`		`- "metric": metric,`
	`57`	`+ "target_metric": metric,`
`58`	`58`	`"node_type": "embedding",`
`59`	`59`	`"search_space": [`
`60`	`60`	`{`