Optional prediction column: the return (#381)

* Make y_pred column optional for estimated and realized performance + tests

* Update data requirements docs

* Fix _list_missing issues + add "run with prediction" tests

* Fix flake8 issues

* Metric classes were not able to deal with optional predictions yet.

Updated the base class and (unfortunately) added asserts to keep mypy from complaining. Didn't want that logic to leak into remove_nans

nannyml/performance_calculation/metrics/base.py

@@ -22,9 +22,9 @@ def __init__(
  self,
  name: str,
  y_true: str,
- y_pred: str,
  components: List[Tuple[str, str]],
  threshold: Threshold,
+ y_pred: Optional[str] = None,
  y_pred_proba: Optional[Union[str, Dict[str, str]]] = None,
  upper_threshold_limit: Optional[float] = None,
  lower_threshold_limit: Optional[float] = None,

nannyml/performance_calculation/metrics/binary_classification.py

@@ -57,8 +57,8 @@ class BinaryClassificationAUROC(Metric):
  def __init__(
  self,
  y_true: str,
- y_pred: str,
  threshold: Threshold,
+ y_pred: Optional[str] = None,
  y_pred_proba: Optional[str] = None,
  **kwargs,
  ):
@@ -97,6 +97,8 @@ def __str__(self):
  def _fit(self, reference_data: pd.DataFrame):
  """Metric _fit implementation on reference data."""
  _list_missing([self.y_true, self.y_pred_proba], list(reference_data.columns))
+ # we don't want to count missing rows for sampling error
+ reference_data = _remove_nans(reference_data, (self.y_true,))
  self._sampling_error_components = auroc_sampling_error_components(
  y_true_reference=reference_data[self.y_true],
  y_pred_proba_reference=reference_data[self.y_pred_proba],
@@ -105,10 +107,10 @@ def _fit(self, reference_data: pd.DataFrame):
  def _calculate(self, data: pd.DataFrame):
  """Redefine to handle NaNs and edge cases."""
  _list_missing([self.y_true, self.y_pred_proba], list(data.columns))
- data = _remove_nans(data, (self.y_true, self.y_pred))
+ data = _remove_nans(data, (self.y_true,))
 
  y_true = data[self.y_true]
- y_pred = data[self.y_pred_proba]
+ y_pred_proba = data[self.y_pred_proba]
 
  if y_true.nunique() <= 1:
  warnings.warn(
@@ -117,7 +119,7 @@ def _calculate(self, data: pd.DataFrame):
  )
  return np.NaN
  else:
- return roc_auc_score(y_true, y_pred)
+ return roc_auc_score(y_true, y_pred_proba)
 
  def _sampling_error(self, data: pd.DataFrame) -> float:
  return auroc_sampling_error(self._sampling_error_components, data)
@@ -133,8 +135,8 @@ class BinaryClassificationAP(Metric):
  def __init__(
  self,
  y_true: str,
- y_pred: str,
  threshold: Threshold,
+ y_pred: Optional[str] = None,
  y_pred_proba: Optional[str] = None,
  **kwargs,
  ):
@@ -174,7 +176,7 @@ def _fit(self, reference_data: pd.DataFrame):
  """Metric _fit implementation on reference data."""
  _list_missing([self.y_true, self.y_pred_proba], list(reference_data.columns))
  # we don't want to count missing rows for sampling error
- reference_data = _remove_nans(reference_data, (self.y_true, self.y_pred))
+ reference_data = _remove_nans(reference_data, (self.y_true,))
 
  if 1 not in reference_data[self.y_true].unique():
  self._sampling_error_components = np.NaN, 0
@@ -187,7 +189,7 @@ def _fit(self, reference_data: pd.DataFrame):
  def _calculate(self, data: pd.DataFrame):
  """Redefine to handle NaNs and edge cases."""
  _list_missing([self.y_true, self.y_pred_proba], list(data.columns))
- data = _remove_nans(data, (self.y_true, self.y_pred))
+ data = _remove_nans(data, (self.y_true,))
 
  y_true = data[self.y_true]
  y_pred_proba = data[self.y_pred_proba]
@@ -259,6 +261,7 @@ def _fit(self, reference_data: pd.DataFrame):
  def _calculate(self, data: pd.DataFrame):
  """Redefine to handle NaNs and edge cases."""
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -335,6 +338,7 @@ def _fit(self, reference_data: pd.DataFrame):
 
  def _calculate(self, data: pd.DataFrame):
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -411,6 +415,7 @@ def _fit(self, reference_data: pd.DataFrame):
 
  def _calculate(self, data: pd.DataFrame):
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -487,6 +492,7 @@ def _fit(self, reference_data: pd.DataFrame):
 
  def _calculate(self, data: pd.DataFrame):
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -564,6 +570,7 @@ def _fit(self, reference_data: pd.DataFrame):
 
  def _calculate(self, data: pd.DataFrame):
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -674,6 +681,7 @@ def _fit(self, reference_data: pd.DataFrame):
 
  def _calculate(self, data: pd.DataFrame):
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -858,6 +866,7 @@ def _fit(self, reference_data: pd.DataFrame):
 
  def _calculate_true_positives(self, data: pd.DataFrame) -> float:
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -882,6 +891,7 @@ def _calculate_true_positives(self, data: pd.DataFrame) -> float:
 
  def _calculate_true_negatives(self, data: pd.DataFrame) -> float:
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -906,6 +916,7 @@ def _calculate_true_negatives(self, data: pd.DataFrame) -> float:
 
  def _calculate_false_positives(self, data: pd.DataFrame) -> float:
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -930,6 +941,7 @@ def _calculate_false_positives(self, data: pd.DataFrame) -> float:
 
  def _calculate_false_negatives(self, data: pd.DataFrame) -> float:
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]

nannyml/performance_calculation/metrics/multiclass_classification.py

@@ -211,6 +211,7 @@ def _calculate(self, data: pd.DataFrame):
  )
 
  _list_missing([self.y_true, self.y_pred], data)
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  labels = sorted(list(self.y_pred_proba.keys()))
@@ -306,6 +307,7 @@ def _calculate(self, data: pd.DataFrame):
  )
 
  _list_missing([self.y_true, self.y_pred], data)
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  labels = sorted(list(self.y_pred_proba.keys()))
@@ -401,6 +403,7 @@ def _calculate(self, data: pd.DataFrame):
  )
 
  _list_missing([self.y_true, self.y_pred], data)
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  labels = sorted(list(self.y_pred_proba.keys()))
@@ -496,6 +499,7 @@ def _calculate(self, data: pd.DataFrame):
  )
 
  _list_missing([self.y_true, self.y_pred], data)
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  labels = sorted(list(self.y_pred_proba.keys()))
@@ -588,6 +592,7 @@ def _fit(self, reference_data: pd.DataFrame):
 
  def _calculate(self, data: pd.DataFrame):
  _list_missing([self.y_true, self.y_pred], data)
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]

nannyml/performance_calculation/metrics/regression.py

@@ -77,6 +77,7 @@ def _fit(self, reference_data: pd.DataFrame):
  def _calculate(self, data: pd.DataFrame):
  """Redefine to handle NaNs and edge cases."""
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -139,6 +140,7 @@ def _fit(self, reference_data: pd.DataFrame):
  def _calculate(self, data: pd.DataFrame):
  """Redefine to handle NaNs and edge cases."""
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -201,6 +203,7 @@ def _fit(self, reference_data: pd.DataFrame):
  def _calculate(self, data: pd.DataFrame):
  """Redefine to handle NaNs and edge cases."""
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -263,6 +266,7 @@ def _fit(self, reference_data: pd.DataFrame):
  def _calculate(self, data: pd.DataFrame):
  """Redefine to handle NaNs and edge cases."""
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -330,6 +334,7 @@ def _fit(self, reference_data: pd.DataFrame):
  def _calculate(self, data: pd.DataFrame):
  """Redefine to handle NaNs and edge cases."""
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]
@@ -392,6 +397,7 @@ def _fit(self, reference_data: pd.DataFrame):
  def _calculate(self, data: pd.DataFrame):
  """Redefine to handle NaNs and edge cases."""
  _list_missing([self.y_true, self.y_pred], list(data.columns))
+ assert self.y_pred
  data = _remove_nans(data, (self.y_true, self.y_pred))
 
  y_true = data[self.y_true]