Update app.py

app.py

@@ -7,12 +7,17 @@ import numpy as np
 from tensorflow.keras.preprocessing.sequence import pad_sequences
 from tensorflow.keras.preprocessing.text import tokenizer_from_json
 
-# Load the Keras model
-def load_model():
-    model = tf.keras.models.load_model('polarisatie_model.h5')
+# Load the Keras model architecture from model.json
+def load_model_architecture(model_json_path='model.json'):
+    with open(model_json_path, 'r', encoding='utf-8') as json_file:
+        model_json = json_file.read()
+    model = tf.keras.models.model_from_json(model_json)
     return model
 
-model = load_model()
+# Load the model weights from polarisatie_model.h5
+def load_model_weights(model, weights_path='polarisatie_model.h5'):
+    model.load_weights(weights_path)
+    return model
 
 # Load the tokenizer from tokenizer.json
 def load_tokenizer(tokenizer_path='tokenizer.json'):
@@ -21,18 +26,14 @@ def load_tokenizer(tokenizer_path='tokenizer.json'):
     tokenizer = tokenizer_from_json(tokenizer_json)
     return tokenizer
 
-tokenizer = load_tokenizer()
-
 # Load max_length from max_length.txt
 def load_max_length(max_length_path='max_length.txt'):
     with open(max_length_path, 'r') as f:
         max_length = int(f.read().strip())
     return max_length
 
-max_length = load_max_length()
-
 # Preprocessing function
-def preprocess(text):
+def preprocess(text, tokenizer, max_length):
     # Tokenize the text
     sequences = tokenizer.texts_to_sequences([text])
     # Pad the sequences
@@ -40,12 +41,12 @@ def preprocess(text):
     return padded
 
 # Prediction function
-def predict_polarization(text):
+def predict_polarization(text, model, tokenizer, max_length):
     if not text.strip():
         return "Voer alstublieft een geldige zin in."
     
     # Preprocess the text
-    processed_text = preprocess(text)
+    processed_text = preprocess(text, tokenizer, max_length)
     
     # Make prediction
     prediction = model.predict(processed_text)
@@ -55,6 +56,12 @@ def predict_polarization(text):
     
     return "Ja" if is_polarizing else "Nee"
 
+# Load all components
+model = load_model_architecture()
+model = load_model_weights(model)
+tokenizer = load_tokenizer()
+max_length = load_max_length()
+
 # Define the Gradio interface using Blocks
 with gr.Blocks() as demo:
     gr.Markdown("# Polarisatie Thermometer")
@@ -67,7 +74,11 @@ with gr.Blocks() as demo:
     
     result = gr.Textbox(label="Is polariserend?", interactive=False)
     
-    evaluate_button.click(fn=predict_polarization, inputs=input_text, outputs=result)
+    evaluate_button.click(
+        fn=lambda text: predict_polarization(text, model, tokenizer, max_length),
+        inputs=input_text,
+        outputs=result
+    )
 
 # Launch the Gradio app
 if __name__ == "__main__":