running app

app.py

@@ -41,22 +41,48 @@ model = load_model('unet_contrails_model.keras', custom_objects={'dice_loss_plus
 model.compile(metrics=metrics)
 
 
-label = np.load('images/human_pixel_masks.npy')
-ash_image = np.load('images/ash_image.npy')[...,4]
-y_pred = model.predict(ash_image.reshape(1,256, 256, 3))
-prediction = np.argmax(y_pred[0], axis=2).reshape(256,256,1)
-intersection = label & prediction
-false_negative = label - intersection
-false_possitive = prediction - intersection 
-color_prediction = np.stack([false_negative*.7, intersection*.7, false_possitive*.7], axis=2).reshape(256,256,3)
-print(intersection)
-
-
-def greet(name):
-    return "Hello " + name + "!!"
-
-iface = gr.Interface(fn=greet, inputs="text", outputs="text")
-iface.launch()
+
+
+
+# def greet(name):
+#     return "Hello " + name + "!!"
+
+# iface = gr.Interface(fn=greet, inputs="text", outputs="text")
+# iface.launch()
+
+'#images/human_pixel_masks.npy'
+def predict(ash_image, model=model):
+    #label = np.load(label_image)
+    ash_image = np.load(ash_image)[...,4]
+    y_pred = model.predict(ash_image.reshape(1,256, 256, 3))
+    prediction = np.argmax(y_pred[0], axis=2).reshape(256,256,1)
+    #intersection = label & prediction
+    #false_negative = label - intersection
+    #false_possitive = prediction - intersection 
+    #color_prediction = np.stack([false_negative*.7, intersection*.7, false_possitive*.7], axis=2).reshape(256,256,3)
+    seg_info = [(prediction, 'contrails')]
+    return(ash_image, seg_info)
+        
+
+
+if __name__ == "__main__":
+    class2hexcolor = {"contrails": "#007fff"}
+
+    with gr.Blocks(title="Contrail Predictions") as demo:
+        gr.Markdown("""<h1><center>Predict Contrails in Satellite Images</center></h1>""")
+        with gr.Row():
+            img_input = gr.Image(type="pil", height=256, width=256, label="Input image")
+            img_output = gr.AnnotatedImage(label="Predictions", height=256, width=256, color_map=class2hexcolor)
+
+        section_btn = gr.Button("Generate Predictions")
+        section_btn.click(partial(predict, model=model), img_input, img_output)
+
+        images_dir = glob(os.path.join(os.getcwd(), "images") + os.sep + "*.png")
+        examples = [i for i in np.random.choice(images_dir, size=1, replace=False)]
+        gr.Examples(examples=examples, inputs=img_input, outputs=img_output)
+
+    demo.launch()
+
 
 
 # first we need to load the model from somewhere - probably by using model.load on a keras file (which could be saved in our huggingface space repo)