Experiment method to plot results

src/aequitas/flow/experiment/experiment.py

@@ -17,7 +17,8 @@
 from ..methods.preprocessing.identity import Identity as PreIdentity
 from ..methods.preprocessing.preprocessing import PreProcessing
 from ..optimization import ObjectiveFunction
-from ..utils import ConfigReader, create_logger, import_object
+from ..utils import ConfigReader, create_logger, import_object, read_results
+from ..plots.pareto import Plot
 
 
 class Experiment:
@@ -143,7 +144,7 @@ def _read_datasets(self):
 
     def run(self) -> None:
         self.logger.info("Beginning Experiment.")
-        exp_folder: Path = Path()
+        self.exp_folder: Path = Path()
         dataset_folder: Path = Path()
         if self.artifacts:
             if self.save_folder is None:
@@ -152,14 +153,14 @@ def run(self) -> None:
             if not self.hash:
                 self.generate_hash()
 
-            exp_folder: Path = self.save_folder / self.hash
-            exp_folder.mkdir(parents=True, exist_ok=True)
-            self.logger.info(f"Saving objects to '{exp_folder.resolve()}'.")
+            self.exp_folder: Path = self.save_folder / self.hash
+            self.exp_folder.mkdir(parents=True, exist_ok=True)
+            self.logger.info(f"Saving objects to '{self.exp_folder.resolve()}'.")
         self._read_datasets()
         for dataset_id, (dataset_name, dataset) in enumerate(self.datasets.items()):
             self.logger.info(f"Using '{dataset_name}'.")
             if self.artifacts:
-                dataset_folder: Path = exp_folder / dataset_name
+                dataset_folder: Path = self.exp_folder / dataset_name
                 dataset_folder.mkdir(parents=True, exist_ok=True)
                 self.logger.debug(
                     f"Saving dataset-related objects to '{dataset_folder.resolve()}'."
@@ -258,3 +259,14 @@ def dt_handler(x):
             json.dumps(self.config, default=dt_handler, sort_keys=True).encode("utf-8")
         ).hexdigest()
         self.logger.debug(f"Hash generated: {self.hash}.")
+
+    def plot_pareto(
+        self,
+        dataset: str,
+        fairness_metric: str = "Predictive Equality",
+        performance_metric: str = "TPR",
+        split: str = "validation",
+    ) -> None:
+        results = read_results(self.exp_folder)
+        plot = Plot(results, dataset, fairness_metric, performance_metric, split=split)
+        return plot.visualize()