handle edge case of no evaluations

dssg · Dec 5, 2023 · 6fc7b6e · 6fc7b6e
1 parent faaa5de
commit 6fc7b6e
Show file tree

Hide file tree

Showing 2 changed files with 9 additions and 4 deletions.
diff --git a/src/triage/component/postmodeling/model_analyzer.py b/src/triage/component/postmodeling/model_analyzer.py
@@ -792,9 +792,13 @@ def plot_precision_recall_curve(self, ax=None, matrix_uuid=None, list_size_upper
 
 
         if pred_df.empty:
-            logging.warning('No predictions were found. Using the evaluations to generate the plot. Zoomed PR-K not supported!')
+            logging.warning(f'No predictions were found for model id {self.model_id} (group: {self.model_group_id}). Using the evaluations to generate the plot. Zoomed PR-K not supported!')
             eval_df = self.get_evaluations(matrix_uuid=matrix_uuid)
 
+            if eval_df.empty: 
+                logging.error('No evaluations were found! Returning empty axes!')
+                return ax
+
             eval_df['perc_points'] = [x.split('_')[0] for x in eval_df['parameter'].tolist()]
             eval_df['perc_points'] = pd.to_numeric(eval_df['perc_points'])
 
@@ -823,6 +827,7 @@ def plot_precision_recall_curve(self, ax=None, matrix_uuid=None, list_size_upper
             )
 
         else:
+            logging.debug(f'Found saved predictions for model id {self.model_id}.(group: {self.model_group_id})')
             k_values = np.arange(0 + pct_step_size, list_size_upper_bound_pct + pct_step_size, pct_step_size)
 
             precisions = list()

diff --git a/src/triage/component/postmodeling/report_generator.py b/src/triage/component/postmodeling/report_generator.py
@@ -187,7 +187,7 @@ def _make_plot_grid(self, plot_type, subplot_width=3, subplot_len=None, sharey=F
                 else:
                     ax = axes[i, j]
 
-                plot_func(ax=ax, **kw)
+                ax = plot_func(ax=ax, **kw)
 
                 if j==0:
                     ax.set_ylabel(f'{train_end_time}')
@@ -212,8 +212,8 @@ def plot_calibration_curves(self):
         """calibration curves for all models"""
         self._make_plot_grid(plot_type='plot_calibration_curve')
 
-    def plot_prk_curves(self):
-        self._make_plot_grid(plot_type='plot_precision_recall_curve')
+    def plot_prk_curves(self, **kw):
+        self._make_plot_grid(plot_type='plot_precision_recall_curve', **kw)
 
     def plot_bias_threshold(self, attribute_name, attribute_values, bias_metric):
         """