added test for interpretation

gradio/interface.py

@@ -266,10 +266,10 @@ class Interface:
             if self.capture_session and self.session is not None:
                 graph, sess = self.session
                 with graph.as_default(), sess.as_default():
-                    interpretation = interpreter(*processed_input).tolist()
+                    interpretation = interpreter(*processed_input)
             else:
                 try:
-                    interpretation = interpreter(*processed_input).tolist()
+                    interpretation = interpreter(*processed_input)
                 except ValueError as exception:
                     if str(exception).endswith("is not an element of this graph."):
                         raise ValueError(strings.en["TF1_ERROR"])

gradio/interpretation.py

@@ -1,6 +1,5 @@
 from gradio.inputs import Image, Textbox
 from gradio.outputs import Label
-from gradio import processing_utils
 from skimage.segmentation import slic
 import numpy as np
 
@@ -8,6 +7,7 @@ expected_types = {
     Image: "numpy",
 }
 
+
 def default(separator=" ", n_segments=20):
     """
     Basic "default" interpretation method that uses "leave-one-out" to explain predictions for

test/test_interpretation.py

@@ -0,0 +1,47 @@
+import unittest
+import gradio.interpretation
+import gradio.test_data
+from gradio.processing_utils import decode_base64_to_image
+from gradio import Interface
+import numpy as np
+
+
+class TestDefault(unittest.TestCase):
+    def setUp(self):
+        self.default_method = gradio.interpretation.default()
+
+    def test_default_text(self):
+        max_word_len = lambda text: max([len(word) for word in text.split(" ")])
+        text_interface = Interface(max_word_len, "textbox", "label")
+        interpretation = self.default_method(text_interface, ["quickest brown fox"])[0]		
+        self.assertGreater(interpretation[0][1], 0)  # Checks to see if the first letter has >0 score.
+        self.assertEqual(interpretation[-1][1], 0)  # Checks to see if the last letter has 0 score.
+
+    def test_default_image(self):
+        max_pixel_value = lambda img: img.max()
+        img_interface = Interface(max_pixel_value, "image", "label")
+        array = np.zeros((100,100))
+        array[0, 0] = 1
+        interpretation = self.default_method(img_interface, [array])[0]
+        self.assertGreater(interpretation[0][0], 0)  # Checks to see if the top-left has >0 score.
+        
+
+class TestCustom(unittest.TestCase):
+    def test_custom_text(self):
+        max_word_len = lambda text: max([len(word) for word in text.split(" ")])
+        custom = lambda text: [(char, 1) for char in text]
+        text_interface = Interface(max_word_len, "textbox", "label", interpretation=custom)
+        result = text_interface.interpret(["quickest brown fox"])[0]
+        self.assertEqual(result[0][1], 1)  # Checks to see if the first letter has score of 1.
+
+    def test_custom_img(self):
+        max_pixel_value = lambda img: img.max()
+        custom = lambda img: img.tolist()
+        img_interface = Interface(max_pixel_value, "image", "label", interpretation=custom)
+        result = img_interface.interpret([gradio.test_data.BASE64_IMAGE])[0]
+        expected_result = np.asarray(decode_base64_to_image(gradio.test_data.BASE64_IMAGE).convert('RGB')).tolist()
+        self.assertEqual(result, expected_result)
+        
+
+if __name__ == '__main__':
+    unittest.main()