Merge remote-tracking branch 'origin/main' into 426-Use-registryfill-…

…to-update-defaults-in-cfg-from-defaults-in-function-signatures

psycop/common/model_evaluation/binary/test_bootstrap_estimates.py

@@ -8,7 +8,7 @@
 
 def test_boostrap_estimates():
     input_df = str_to_df(
-        """id,y,y_hat,y_pred,
+        """id,y,y_hat_prob,y_hat,
         1,1,1.0,1,
         1,1,1.0,1,
         1,1,1.0,1,
@@ -26,7 +26,7 @@ def test_boostrap_estimates():
         n_bootstraps=1,
         ci_width=0.95,
         input_1=input_df["y"],
-        input_2=input_df["y_hat"],
+        input_2=input_df["y_hat_prob"],
     )
 
     assert auroc_df_with_ci["ci"] == (1.0, 1.0)
@@ -36,7 +36,7 @@ def test_boostrap_estimates():
         n_bootstraps=5,
         ci_width=0.95,
         input_1=input_df["y"],
-        input_2=input_df["y_pred"],
+        input_2=input_df["y_hat"],
     )
 
     assert sensitivity_df_with_ci["ci"] == (1.0, 1.0)

psycop/common/model_training_v2/test_pipeline.py

@@ -18,7 +18,7 @@
     TerminalLogger,
 )
 from psycop.common.model_training_v2.trainer.base_dataloader import BaselineDataLoader
-from psycop.common.model_training_v2.trainer.cross_validator import (
+from psycop.common.model_training_v2.trainer.cross_validator_trainer import (
     CrossValidatorTrainer,
 )
 from psycop.common.model_training_v2.trainer.preprocessing.pipeline import (
@@ -30,12 +30,12 @@
 from psycop.common.model_training_v2.trainer.split_trainer import (
     SplitTrainer,
 )
-from psycop.common.model_training_v2.trainer.task.binary_classification.binary_classification import (
-    BinaryClassification,
-)
 from psycop.common.model_training_v2.trainer.task.binary_classification.binary_classification_pipeline import (
     BinaryClassificationPipeline,
 )
+from psycop.common.model_training_v2.trainer.task.binary_classification.binary_classification_task import (
+    BinaryClassificationTask,
+)
 from psycop.common.model_training_v2.trainer.task.binary_classification.binary_metrics import (
     BinaryAUROC,
 )
@@ -80,7 +80,7 @@ def test_v2_train_model_pipeline(tmpdir: Path):
             preprocessing_pipeline=BaselinePreprocessingPipeline(
                 AgeFilter(min_age=4, max_age=99, age_col_name="pred_age"),
             ),
-            task=BinaryClassification(
+            task=BinaryClassificationTask(
                 pred_time_uuid_col_name="pred_time_uuid",
                 task_pipe=BinaryClassificationPipeline(
                     sklearn_pipe=Pipeline([logistic_regression_step()]),
@@ -117,7 +117,7 @@ def test_v2_crossval_model_pipeline(tmpdir: Path):
             preprocessing_pipeline=BaselinePreprocessingPipeline(
                 AgeFilter(min_age=4, max_age=99, age_col_name="pred_age"),
             ),
-            task=BinaryClassification(
+            task=BinaryClassificationTask(
                 pred_time_uuid_col_name="pred_time_uuid",
                 task_pipe=BinaryClassificationPipeline(
                     sklearn_pipe=Pipeline([logistic_regression_step()]),

psycop/common/model_training_v2/trainer/cross_validator.py → psycop/common/model_training_v2/trainer/cross_validator_trainer.py

@@ -12,7 +12,7 @@
 from psycop.common.model_training_v2.trainer.preprocessing.pipeline import (
     PreprocessingPipeline,
 )
-from psycop.common.model_training_v2.trainer.task.base_metric import BaseMetric
+from psycop.common.model_training_v2.trainer.task.base_metric import BaselineMetric
 from psycop.common.model_training_v2.trainer.task.base_task import BaselineTask
 
 
@@ -24,7 +24,7 @@ def __init__(
         outcome_col_name: str,
         preprocessing_pipeline: PreprocessingPipeline,
         task: BaselineTask,
-        metric: BaseMetric,
+        metric: BaselineMetric,
         logger: BaselineLogger,
         n_splits: int = 5,
         group_col_name: str = "dw_ek_borger",
@@ -97,12 +97,12 @@ def train(self) -> TrainingResult:
 
             training_data_preprocessed.loc[
                 val_idxs,
-                "oof_y_hat_probs",
+                "oof_y_hat_prob",
             ] = oof_y_hat_prob.to_list()  # type: ignore
 
         main_metric = self.metric.calculate(
             y=training_data_preprocessed[self.outcome_col_name],
-            y_hat_prob=training_data_preprocessed["oof_y_hat_probs"],
+            y_hat_prob=training_data_preprocessed["oof_y_hat_prob"],
         )
         self.logger.log_metric(main_metric)
 

psycop/common/model_training_v2/trainer/split_trainer.py

@@ -19,7 +19,7 @@
 from psycop.common.model_training_v2.trainer.preprocessing.pipeline import (
     PreprocessingPipeline,
 )
-from psycop.common.model_training_v2.trainer.task.base_metric import BaseMetric
+from psycop.common.model_training_v2.trainer.task.base_metric import BaselineMetric
 from psycop.common.model_training_v2.trainer.task.base_task import (
     BaselineTask,
 )
@@ -35,7 +35,7 @@ def __init__(
         validation_outcome_col_name: str,
         preprocessing_pipeline: PreprocessingPipeline,
         task: BaselineTask,
-        metric: BaseMetric,
+        metric: BaselineMetric,
         logger: BaselineLogger,
     ):
         self.training_data = training_data.load()

psycop/common/model_training_v2/trainer/task/base_metric.py

@@ -10,11 +10,11 @@ class CalculatedMetric:
     value: float
 
 
-PredProbaSeries = pd.Series  # name should be "y_hat_probs", series of floats
+PredProbaSeries = pd.Series  # name should be "y_hat_prob", series of floats
 
 
 @runtime_checkable
-class BaseMetric(Protocol):
+class BaselineMetric(Protocol):
     def calculate(
         self,
         y: pd.Series,  # type: ignore

psycop/common/model_training_v2/trainer/task/binary_classification/binary_classification_pipeline.py

@@ -17,4 +17,4 @@ def predict_proba(self, x: pd.DataFrame) -> PredProbaSeries:
         """Returns the predicted probabilities of the `1`
         class"""
         pred_probs = self.pipe.predict_proba(x)[:, 1]
-        return pd.Series(pred_probs, name="y_hat_probs")
+        return pd.Series(pred_probs, name="y_hat_prob")

psycop/common/model_training_v2/trainer/task/binary_classification/binary_classification.py → psycop/common/model_training_v2/trainer/task/binary_classification/binary_classification_task.py

@@ -22,7 +22,7 @@ def polarsframe_to_series(polarsframe: PolarsFrame) -> pl.Series:
 
 
 @BaselineRegistry.tasks.register("binary_classification")
-class BinaryClassification(BaselineTask):
+class BinaryClassificationTask(BaselineTask):
     def __init__(
         self,
         task_pipe: BinaryClassificationPipeline,

psycop/common/model_training_v2/trainer/task/binary_classification/binary_metrics/base_binary_metric.py

@@ -2,7 +2,7 @@
 
 from typing import TYPE_CHECKING
 
-from psycop.common.model_training_v2.trainer.task.base_metric import BaseMetric
+from psycop.common.model_training_v2.trainer.task.base_metric import BaselineMetric
 
 if TYPE_CHECKING:
     import pandas as pd
@@ -13,7 +13,7 @@
     )
 
 
-class BinaryMetric(BaseMetric):
+class BinaryMetric(BaselineMetric):
     def calculate(
         self,
         y: pd.Series,  # type: ignore

psycop/common/model_training_v2/trainer/task/binary_classification/binary_metrics/test_binary_auroc.py

@@ -16,7 +16,7 @@
 
 
 @pytest.mark.parametrize(
-    ("y_true", "y_pred", "expected"),
+    ("y", "y_hat_prob", "expected"),
     [
         (
             pd.Series([1, 1, 0, 0]),
@@ -30,7 +30,7 @@
         ),
     ],
 )
-def test_binary_auroc(y_true: pd.Series[int], y_pred: PredProbaSeries, expected: float):
+def test_binary_auroc(y: pd.Series[int], y_hat_prob: PredProbaSeries, expected: float):
     auroc = BinaryAUROC()
-    calculated_metric = auroc.calculate(y=y_true, y_hat_prob=y_pred)
+    calculated_metric = auroc.calculate(y=y, y_hat_prob=y_hat_prob)
     assert calculated_metric.value == expected

psycop/common/model_training_v2/trainer/task/test_binary_classification.py → psycop/common/model_training_v2/trainer/task/binary_classification/test_binary_classification.py

@@ -3,12 +3,12 @@
 from pandas.testing import assert_series_equal
 from sklearn.pipeline import Pipeline
 
-from psycop.common.model_training_v2.trainer.task.binary_classification.binary_classification import (
-    BinaryClassification,
-)
 from psycop.common.model_training_v2.trainer.task.binary_classification.binary_classification_pipeline import (
     BinaryClassificationPipeline,
 )
+from psycop.common.model_training_v2.trainer.task.binary_classification.binary_classification_task import (
+    BinaryClassificationTask,
+)
 from psycop.common.model_training_v2.trainer.task.binary_classification.binary_metrics.binary_auroc import (
     BinaryAUROC,
 )
@@ -38,7 +38,7 @@ def test_binary_classification(
     y: pd.DataFrame,
     main_metric_expected: float,
 ):
-    binary_classification_problem = BinaryClassification(
+    binary_classification_problem = BinaryClassificationTask(
         task_pipe=pipe,
         pred_time_uuid_col_name="uuid",
     )

psycop/common/model_training_v2/trainer/task/binary_classification/test_binary_classification_pipeline.py

@@ -28,6 +28,6 @@ def test_binary_classification_pipeline(
     pipeline = BinaryClassificationPipeline(sklearn_pipe=pipe)
     pipeline.fit(x=x, y=y)
 
-    y_hat_probs = pipeline.predict_proba(x=x)
-    assert isinstance(y_hat_probs, pd.Series)
-    assert y_hat_probs.name == "y_hat_probs"
+    y_hat_prob = pipeline.predict_proba(x=x)
+    assert isinstance(y_hat_prob, pd.Series)
+    assert y_hat_prob.name == "y_hat_prob"

psycop/common/model_training_v2/trainer/task/multilabel_classification/multiclass_classification_pipeline.py → psycop/common/model_training_v2/trainer/task/multilabel_classification/multilabel_classification_pipeline.py

@@ -9,7 +9,7 @@
 )
 
 
-class MulticlassClassificationPipeline(Protocol):
+class MultilabelClassificationPipeline(Protocol):
     def __init__(self, steps: Sequence[ModelStep]) -> None:
         ...
 

psycop/common/model_training_v2/trainer/task/multilabel_classification/multilabel_classification.py → psycop/common/model_training_v2/trainer/task/multilabel_classification/multilabel_classification_task.py

@@ -5,18 +5,18 @@
 from psycop.common.model_training_v2.trainer.task.base_task import (
     BaselineTask,
 )
-from psycop.common.model_training_v2.trainer.task.multilabel_classification.multiclass_classification_pipeline import (
-    MulticlassClassificationPipeline,
+from psycop.common.model_training_v2.trainer.task.multilabel_classification.multilabel_classification_pipeline import (
+    MultilabelClassificationPipeline,
 )
-from psycop.common.model_training_v2.trainer.task.multilabel_classification.multilabel_metrics.base import (
+from psycop.common.model_training_v2.trainer.task.multilabel_classification.multilabel_metrics.base_multilabel_metric import (
     MultilabelMetric,
 )
 
 
-class MultilabelClassification(BaselineTask):
+class MultilabelClassificationTask(BaselineTask):
     def __init__(
         self,
-        pipe: MulticlassClassificationPipeline,
+        pipe: MultilabelClassificationPipeline,
         main_metric: MultilabelMetric,
         supplementary_metrics: Sequence[MultilabelMetric] | None = None,
     ):

psycop/common/model_training_v2/trainer/task/multilabel_classification/multilabel_metrics/base_multilabel_metric.py

@@ -1,13 +1,13 @@
 import pandas as pd
 
 from psycop.common.model_training_v2.trainer.task.base_metric import (
-    BaseMetric,
+    BaselineMetric,
     CalculatedMetric,
     PredProbaSeries,
 )
 
 
-class MultilabelMetric(BaseMetric):
+class MultilabelMetric(BaselineMetric):
     def calculate(
         self,
         y: pd.Series,  # type: ignore