Added progressbar to tdqm

quantus/helpers/pytorch_model.py

@@ -35,7 +35,11 @@ def predict(self, x, **kwargs):
             return pred.cpu().numpy()
 
     def shape_input(
-        self, x: np.array, shape: Tuple[int, ...], channel_first: Optional[bool] = None, batch: bool = False
+        self,
+        x: np.array,
+        shape: Tuple[int, ...],
+        channel_first: Optional[bool] = None,
+        batch: bool = False,
     ):
         """
         Reshape input into model expected input.

quantus/helpers/tf_model.py

@@ -40,7 +40,11 @@ def predict(self, x, **kwargs):
         return new_model(x, training=False).numpy()
 
     def shape_input(
-        self, x: np.array, shape: Tuple[int, ...], channel_first: Optional[bool] = None, batch: bool = False
+        self,
+        x: np.array,
+        shape: Tuple[int, ...],
+        channel_first: Optional[bool] = None,
+        batch: bool = False,
     ):
         """
         Reshape input into model expected input.

quantus/helpers/warn_func.py

@@ -70,5 +70,6 @@ def deprecation_warnings(kwargs: dict = {}) -> None:
         text = "argument 'nr_channels' is deprecated and will be removed in future versions.\n"
     if "max_steps_per_input" in kwargs:
         text = "argument 'max_steps_per_input' is deprecated and will be removed in future versions.\n"
-
-    if text != '\n': print(text)
+
+    if text != "\n":
+        print(text)

quantus/metrics/axiomatic_metrics.py

@@ -200,7 +200,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 
@@ -422,7 +426,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 
@@ -640,7 +648,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 

quantus/metrics/base.py

@@ -115,7 +115,6 @@ def interpret_scores(self) -> None:
 
         """
         print(self.__init__.__doc__.split(".")[1].split("References")[0])
-        # print(self.__call__.__doc__.split("callable.")[1].split("Parameters")[0])
 
     @property
     def get_params(self) -> dict:

quantus/metrics/complexity_metrics.py

@@ -184,7 +184,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 
@@ -383,7 +387,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 
@@ -575,7 +583,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 

quantus/metrics/faithfulness_metrics.py

@@ -211,7 +211,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 
@@ -441,7 +445,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 
@@ -672,7 +680,9 @@ def __call__(
             iterator = enumerate(zip(x_batch_s, y_batch, a_batch))
         else:
             iterator = tqdm(
-                enumerate(zip(x_batch_s, y_batch, a_batch)), total=len(x_batch_s)
+                enumerate(zip(x_batch_s, y_batch, a_batch)),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         for ix, (x, y, a) in iterator:
@@ -920,7 +930,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 
@@ -1157,7 +1171,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 
@@ -1404,7 +1422,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 
@@ -1650,7 +1672,9 @@ def __call__(
             iterator = enumerate(zip(x_batch_s, y_batch, a_batch))
         else:
             iterator = tqdm(
-                enumerate(zip(x_batch_s, y_batch, a_batch)), total=len(x_batch_s)
+                enumerate(zip(x_batch_s, y_batch, a_batch)),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         for sample, (x, y, a) in iterator:
@@ -1753,7 +1777,9 @@ def __call__(
             self.last_results[sample] = sub_results
 
         if self.return_aggregate:
-            print("A 'return_aggregate' functionality is not implemented for this metric.")
+            print(
+                "A 'return_aggregate' functionality is not implemented for this metric."
+            )
 
         self.all_results.append(self.last_results)
 
@@ -1946,7 +1972,9 @@ def __call__(
             iterator = enumerate(zip(x_batch_s, y_batch, a_batch))
         else:
             iterator = tqdm(
-                enumerate(zip(x_batch_s, y_batch, a_batch)), total=len(x_batch_s)
+                enumerate(zip(x_batch_s, y_batch, a_batch)),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         for sample, (x, y, a) in iterator:
@@ -2035,7 +2063,9 @@ def __call__(
             self.last_results[sample] = sub_results
 
         if self.return_aggregate:
-            print("A 'return_aggregate' functionality is not implemented for this metric.")
+            print(
+                "A 'return_aggregate' functionality is not implemented for this metric."
+            )
 
         self.all_results.append(self.last_results)
 
@@ -2256,7 +2286,9 @@ def __call__(
             iterator = enumerate(zip(x_batch_s, y_batch, a_batch))
         else:
             iterator = tqdm(
-                enumerate(zip(x_batch_s, y_batch, a_batch)), total=len(x_batch_s)
+                enumerate(zip(x_batch_s, y_batch, a_batch)),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         for sample, (x, y, a) in iterator:
@@ -2522,7 +2554,11 @@ def __call__(
         if not self.display_progressbar:
             iterator = zip(x_batch_s, y_batch, a_batch)
         else:
-            iterator = tqdm(zip(x_batch_s, y_batch, a_batch), total=len(x_batch_s))
+            iterator = tqdm(
+                zip(x_batch_s, y_batch, a_batch),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
+            )
 
         for x, y, a in iterator:
 
@@ -2779,7 +2815,9 @@ def __call__(
             iterator = enumerate(zip(x_batch_s, y_batch, a_batch))
         else:
             iterator = tqdm(
-                enumerate(zip(x_batch_s, y_batch, a_batch)), total=len(x_batch_s)
+                enumerate(zip(x_batch_s, y_batch, a_batch)),
+                total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         self.last_results = {str(k): 0 for k in self.percentages}
@@ -2818,7 +2856,9 @@ def __call__(
 
         # Calculate accuracy for every number of most important pixels removed.
         if self.return_aggregate:
-            print("A 'return_aggregate' functionality is not implemented for this metric.")
+            print(
+                "A 'return_aggregate' functionality is not implemented for this metric."
+            )
         for k in self.last_results:
             self.last_results[k] = self.last_results[k] / len(x_batch_s)
 
@@ -2974,7 +3014,9 @@ def __call__(
         a_sim_matrix[dist_matrix <= self.threshold] = 1
 
         # Predict on input.
-        x_input = model.shape_input(x_batch, x_batch[0].shape, channel_first=True, batch=True)
+        x_input = model.shape_input(
+            x_batch, x_batch[0].shape, channel_first=True, batch=True
+        )
         y_pred_classes = np.argmax(
             model.predict(x_input, softmax_act=True, **self.kwargs), axis=1
         ).flatten()
@@ -2986,6 +3028,7 @@ def __call__(
             iterator = tqdm(
                 enumerate(zip(x_batch_s, y_batch, a_batch, a_sim_matrix)),
                 total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         for ix, (x, y, a, a_sim) in iterator:
@@ -3003,7 +3046,9 @@ def __call__(
                 )
 
         if self.return_aggregate:
-            self.last_results = [self.aggregate_func(self.last_results) / len(self.last_results)]
+            self.last_results = [
+                self.aggregate_func(self.last_results) / len(self.last_results)
+            ]
 
         self.all_results.append(self.last_results)
 

quantus/metrics/localisation_metrics.py

@@ -189,6 +189,7 @@ def __call__(
             iterator = tqdm(
                 enumerate(zip(x_batch_s, y_batch, a_batch, s_batch)),
                 total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         for ix, (x, y, a, s) in iterator:
@@ -214,7 +215,9 @@ def __call__(
             else:
                 hit = bool(s[max_index])
 
-            self.last_results.append(hit) # ratio = np.sum(binary_mask) / float(binary_mask.shape[0] * binary_mask.shape[1])
+            self.last_results.append(
+                hit
+            )  # ratio = np.sum(binary_mask) / float(binary_mask.shape[0] * binary_mask.shape[1])
 
         if self.return_aggregate:
             self.last_results = [self.aggregate_func(self.last_results)]
@@ -403,6 +406,7 @@ def __call__(
             iterator = tqdm(
                 enumerate(zip(x_batch_s, y_batch, a_batch, s_batch)),
                 total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         for ix, (x, y, a, s) in iterator:
@@ -638,6 +642,7 @@ def __call__(
             iterator = tqdm(
                 enumerate(zip(x_batch_s, y_batch, a_batch, s_batch)),
                 total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         for ix, (x, y, a, s) in iterator:
@@ -846,6 +851,7 @@ def __call__(
             iterator = tqdm(
                 enumerate(zip(x_batch_s, y_batch, a_batch, s_batch)),
                 total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         for ix, (x, y, a, s) in iterator:
@@ -1052,6 +1058,7 @@ def __call__(
             iterator = tqdm(
                 enumerate(zip(x_batch_s, y_batch, a_batch, s_batch)),
                 total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         for ix, (x, y, a, s) in iterator:
@@ -1249,6 +1256,7 @@ def __call__(
             iterator = tqdm(
                 enumerate(zip(x_batch_s, y_batch, a_batch, s_batch)),
                 total=len(x_batch_s),
+                desc=f"Evaluation of {self.__class__.__name__} metric.",
             )
 
         for ix, (x, y, a, s) in iterator: