Update app.py

app.py

@@ -1,5 +1,4 @@
 import gradio as gr
-from tensorflow import keras
 import pandas as pd
 import tensorflow as tf
 import nltk
@@ -10,6 +9,7 @@ from nltk.tokenize import word_tokenize
 from tensorflow.keras.preprocessing.text import Tokenizer
 from tensorflow.keras.preprocessing.sequence import pad_sequences
 
+# Download and load necessary resources
 import spacy.cli
 spacy.cli.download("en_core_web_sm")
 nltk.download('punkt_tab')
@@ -17,17 +17,20 @@ nltk.download('stopwords')
 stop_words = set(stopwords.words('english'))
 nlp = spacy.load('en_core_web_sm')
 
-# Available backend options are: "jax", "torch", "tensorflow".
-import os
-os.environ["KERAS_BACKEND"] = "jax"
+# Download the model file from Hugging Face
+import requests
+model_url = "https://huggingface.co/Zmorell/HIPA_2/resolve/main/saved_keras_model.keras"
+local_model_path = "saved_keras_model.keras"
 
-# Ensure the necessary libraries are correctly imported
-import keras
+response = requests.get(model_url)
+with open(local_model_path, 'wb') as f:
+    f.write(response.content)
 
-# Load the model from the Hugging Face repository
-model_path = "https://huggingface.co/Zmorell/HIPA_2/resolve/main/saved_keras_model.keras"
-model = tf.keras.models.load_model(model_path)
-print(f"Model loaded from {model_path}")
+print(f"Model downloaded to {local_model_path}")
+
+# Load the downloaded model
+model = tf.keras.models.load_model(local_model_path)
+print(f"Model loaded from {local_model_path}")
 
 def preprocess_text(text):
     text = re.sub(r'[^a-zA-Z0-9\s]', '', text)  # Only remove non-alphanumeric characters except spaces
@@ -44,9 +47,11 @@ def preprocess_text(text):
 def predict(text):
     inputs = preprocess_text(text)
     # Ensure the input shape matches what the model expects
-    inputs = tf.convert_to_tensor([inputs])
-    outputs = model(inputs)
-    return "This text is a violation = " + str(outputs[0][0].numpy())
+    inputs = tokenizer.texts_to_sequences([inputs])
+    inputs = pad_sequences(inputs, maxlen=1000, padding='post')
+    outputs = model.predict(inputs)
+    return f"This text is a violation = {outputs[0][0]:.2f}"
 
+# Set up the Gradio interface
 demo = gr.Interface(fn=predict, inputs="text", outputs="text")
-demo.launch()
+demo.launch()