Create app.py

app.py

@@ -0,0 +1,44 @@
+from transformers import LongformerTokenizerFast, LongformerForSequenceClassification
+import torch
+import gradio as gr
+import requests
+from bs4 import BeautifulSoup
+
+
+def get_text_from_url(url):
+
+    headers = {
+        'Accept-Language': 'en-US,en;q=0.9',
+    }
+    response = requests.get(url, headers=headers)
+    if response.status_code == 200:
+        soup = BeautifulSoup(response.content, 'html.parser')
+
+        texto = soup.get_text()
+
+        return texto
+    else:
+
+        print("Error al obtener la página:", response.status_code)
+        return None
+
+classification_model_checkpoint = 'FrancoMartino/privacyPolicies_classification'
+classification_tokenizer = LongformerTokenizerFast.from_pretrained(classification_model_checkpoint)
+classification_model = LongformerForSequenceClassification.from_pretrained(classification_model_checkpoint)
+
+def predict(url):
+    text = get_text_from_url(url)
+    inputs = classification_tokenizer(text, return_tensors="pt", truncation=True, padding=True, max_length=4096)
+    with torch.no_grad():
+        logits = classification_model(**inputs).logits
+    probabilities = torch.softmax(logits, dim=1)
+    prediction = probabilities[:,1].item()
+    return {'Risk Indicator': prediction}
+
+examples_urls = [
+    ["https://help.instagram.com/155833707900388"],
+    ["https://www.apple.com/legal/privacy/en-ww/"],
+]
+
+interface = gr.Interface(fn=predict, inputs="text",examples=examples_urls, outputs="label", title="Privacy Policy Risk Indicator", description="Enter a privacy policy URL to calculate risk.")
+interface.launch()