make label a required argument (Fixes #131)

dashboard/callbacks.py

@@ -163,10 +163,10 @@ def global_store_i(method_sel, model_path, image_test, labels=list(range(2)),
     elif method_sel == "KernelSHAP":
         relevances = dianna.explain_image(
             model_path, image_test,
+            labels=labels,
             method=method_sel, nsamples=n_samples,
             background=background, n_segments=n_segments, sigma=sigma,
             axis_labels=axis_labels)
-
     else:
         relevances = dianna.explain_image(
             model_path, image_test * 256, 'LIME',

dianna/__init__.py

@@ -33,7 +33,7 @@
 __version__ = "0.6.0"
 
 
-def explain_image(model_or_function, input_data, method, labels=(1,), **kwargs):
+def explain_image(model_or_function, input_data, method, labels, **kwargs):
     """
     Explain an image (input_data) given a model and a chosen method.
 
@@ -42,7 +42,7 @@ def explain_image(model_or_function, input_data, method, labels=(1,), **kwargs):
                                              the path to a ONNX model on disk.
         input_data (np.ndarray): Image data to be explained
         method (string): One of the supported methods: RISE, LIME or KernelSHAP
-        labels (tuple): Labels to be explained
+        labels (Iterable(int)): Labels to be explained
 
     Returns:
         One heatmap (2D array) per class.
@@ -56,7 +56,7 @@ def explain_image(model_or_function, input_data, method, labels=(1,), **kwargs):
     return explainer.explain(model_or_function, input_data, labels, **explain_image_kwargs)
 
 
-def explain_text(model_or_function, input_text, tokenizer, method, labels=(1,), **kwargs):
+def explain_text(model_or_function, input_text, tokenizer, method, labels, **kwargs):
     """
     Explain text (input_text) given a model and a chosen method.
 
@@ -66,7 +66,7 @@ def explain_text(model_or_function, input_text, tokenizer, method, labels=(1,),
         input_text (string): Text to be explained
         tokenizer : Tokenizer class with tokenize and convert_tokens_to_string methods, and mask_token attribute
         method (string): One of the supported methods: RISE or LIME
-        labels (tuple): Labels to be explained
+        labels (Iterable(int)): Labels to be explained
 
     Returns:
         List of (word, index of word in raw text, importance for target class) tuples.

dianna/methods/kernelshap.py

@@ -62,7 +62,7 @@ def explain(
         self,
         model,
         input_data,
-        labels=(0,),
+        labels,
         nsamples="auto",
         background=None,
         n_segments=100,
@@ -84,7 +84,7 @@ def explain(
                                      example. The input dimension must be
                                      [batch, height, width, color_channels] or
                                      [batch, color_channels, height, width] (see axis_labels)
-            labels (tuple): Indices of classes to be explained
+            labels (Iterable(int)): Indices of classes to be explained
             nsamples ("auto" or int): Number of times to re-evaluate the model when
                                       explaining each prediction. More samples lead
                                       to lower variance estimates of the SHAP values.

dianna/methods/lime.py

@@ -52,7 +52,7 @@ def __init__(self,
     def explain(self,
                 model_or_function,
                 input_text,
-                labels=(0,),
+                labels,
                 tokenizer=None,
                 top_labels=None,
                 num_features=10,
@@ -67,7 +67,7 @@ def explain(self,
                                                  the path to a ONNX model on disk.
             tokenizer : Tokenizer class with tokenize and convert_tokens_to_string methods, and mask_token attribute
             input_text (np.ndarray): Data to be explained
-            labels ([int], optional): Iterable of indices of class to be explained
+            labels (Iterable(int)): Iterable of indices of class to be explained
 
         Other keyword arguments: see the LIME documentation for LimeTextExplainer.explain_instance:
         https://lime-ml.readthedocs.io/en/latest/lime.html#lime.lime_text.LimeTextExplainer.explain_instance.
@@ -145,7 +145,7 @@ def __init__(self,
     def explain(self,
                 model_or_function,
                 input_data,
-                labels=(1,),
+                labels,
                 top_labels=None,
                 num_features=10,
                 num_samples=5000,
@@ -161,7 +161,7 @@ def explain(self,
                                                  the path to a ONNX model on disk.
             input_data (np.ndarray): Data to be explained. Must be an "RGB image", i.e. with values in
                                      the [0,255] range.
-            labels (tuple): Indices of classes to be explained
+            labels (Iterable(int)): Indices of classes to be explained
         Other keyword arguments: see the LIME documentation for LimeImageExplainer.explain_instance and
         ImageExplanation.get_image_and_mask:
 

dianna/methods/rise.py

@@ -44,7 +44,7 @@ def __init__(self, n_masks=1000, feature_res=8, p_keep=None,
         self.masks = None
         self.predictions = None
 
-    def explain(self, model_or_function, input_text, labels=(0,), tokenizer=None, batch_size=100):
+    def explain(self, model_or_function, input_text, labels, tokenizer=None, batch_size=100):
         """Runs the RISE explainer on text.
 
            The model will be called with masked versions of the input text.
@@ -54,7 +54,7 @@ def explain(self, model_or_function, input_text, labels=(0,), tokenizer=None, ba
                                                  the path to a ONNX model on disk.
             input_text (np.ndarray): Text to be explained
             tokenizer: Tokenizer class with tokenize and convert_tokens_to_string methods, and mask_token attribute
-            labels (list(int)): Labels to be explained
+            labels (Iterable(int)): Labels to be explained
             batch_size (int): Batch size to use for running the model.
 
         Returns:
@@ -147,7 +147,7 @@ def __init__(self, n_masks=1000, feature_res=8, p_keep=None,
         self.predictions = None
         self.axis_labels = axis_labels if axis_labels is not None else []
 
-    def explain(self, model_or_function, input_data, labels=None, batch_size=100):
+    def explain(self, model_or_function, input_data, labels, batch_size=100):
         """Runs the RISE explainer on images.
 
            The model will be called with masked images,
@@ -158,7 +158,7 @@ def explain(self, model_or_function, input_data, labels=None, batch_size=100):
                                                  the path to a ONNX model on disk.
             input_data (np.ndarray): Image to be explained
             batch_size (int): Batch size to use for running the model.
-            labels (tuple): Labels to be explained
+            labels (Iterable(int)): Labels to be explained
 
         Returns:
             Explanation heatmap for each class (np.ndarray).

tests/test_common_usage.py

@@ -1,14 +1,15 @@
 import numpy as np
+
 import dianna
 import dianna.visualization
 from tests.utils import run_model
 
-
 input_data = np.random.random((224, 224, 3))
 axis_labels = {-1: 'channels'}
+labels = [0, 1]
 
 
 def test_common_RISE_pipeline():  # noqa: N802 ignore case
-    heatmap = dianna.explain_image(run_model, input_data, method="RISE", axis_labels=axis_labels)[0]
+    heatmap = dianna.explain_image(run_model, input_data, "RISE",  labels, axis_labels=axis_labels)[0]
     dianna.visualization.plot_image(heatmap, show_plot=False)
     dianna.visualization.plot_image(heatmap, original_data=input_data[0], show_plot=False)

tests/test_kernelshap.py

@@ -53,9 +53,11 @@ def test_shap_explain_image(self):
         onnx_model_path = "./tests/test_data/mnist_model.onnx"
         n_segments = 50
         explainer = KERNELSHAPImage()
+        labels = [0]
         shap_values, _ = explainer.explain(
             onnx_model_path,
             input_data,
+            labels,
             nsamples=1000,
             background=0,
             n_segments=n_segments,

tests/test_lime.py

@@ -15,9 +15,10 @@ def test_lime_function(self):
         np.random.seed(42)
         input_data = np.random.random((224, 224, 3))
         heatmap_expected = np.load('tests/test_data/heatmap_lime_function.npy')
+        labels = [1]
 
         explainer = LIMEImage(random_state=42)
-        heatmap = explainer.explain(run_model, input_data, num_samples=100)
+        heatmap = explainer.explain(run_model, input_data, labels, num_samples=100)
 
         assert heatmap[0].shape == input_data.shape[:2]
         assert np.allclose(heatmap, heatmap_expected, atol=1e-5)
@@ -27,10 +28,11 @@ def test_lime_filename(self):
         np.random.seed(42)
         model_filename = 'tests/test_data/mnist_model.onnx'
         input_data = generate_data(batch_size=1)[0].astype(np.float32)
-        labels = ('channels', 'y', 'x')
+        axis_labels = ('channels', 'y', 'x')
+        labels = [1]
 
-        heatmap = dianna.explain_image(model_filename, input_data, method="LIME", random_state=42,
-                                       axis_labels=labels)
+        heatmap = dianna.explain_image(model_filename, input_data, method="LIME", labels=labels, random_state=42,
+                                       axis_labels=axis_labels)
 
         heatmap_expected = np.load('tests/test_data/heatmap_lime_filename.npy')
         assert heatmap[0].shape == input_data[0].shape

tests/test_rise.py

@@ -16,22 +16,25 @@ def test_rise_function(self):
         """Test if rise runs and outputs the correct shape given some data and a model function."""
         input_data = np.random.random((224, 224, 3))
         axis_labels = ['y', 'x', 'channels']
-
-        heatmaps = dianna.explain_image(run_model, input_data, method="RISE", axis_labels=axis_labels, n_masks=200, p_keep=.5)
+        labels = [1]
         heatmaps_expected = np.load('tests/test_data/heatmap_rise_function.npy')
 
+        heatmaps = dianna.explain_image(run_model, input_data, "RISE", labels, axis_labels=axis_labels, n_masks=200, p_keep=.5)
+
         assert heatmaps[0].shape == input_data.shape[:2]
         assert np.allclose(heatmaps, heatmaps_expected, atol=1e-5)
 
     def test_rise_filename(self):
         """Test if rise runs and outputs the correct shape given some data and a model file."""
         model_filename = 'tests/test_data/mnist_model.onnx'
         input_data = generate_data(batch_size=1).astype(np.float32)[0]
-
-        heatmaps = dianna.explain_image(model_filename, input_data, method="RISE", n_masks=200, p_keep=.5)
         heatmaps_expected = np.load('tests/test_data/heatmap_rise_filename.npy')
+        labels = [1]
+
+        heatmaps = dianna.explain_image(model_filename, input_data, "RISE", labels, n_masks=200, p_keep=.5)
 
         assert heatmaps[0].shape == input_data.shape[1:]
+        print(heatmaps_expected.shape)
         assert np.allclose(heatmaps, heatmaps_expected, atol=1e-5)
 
     def test_rise_determine_p_keep_for_images(self):