Update app.py

app.py

@@ -1,11 +1,57 @@
-import traceback  # Asegúrate de tener esto al inicio de tu script
+# app.py
 
+import gradio as gr
+import torch
+import numpy as np
+import matplotlib.pyplot as plt
+import base64
+import io
+from fastai.vision.all import *
+import tensorflow as tf
+from tensorflow import keras
+import zipfile
+import os
+import traceback
+
+# Cargar modelo TensorFlow ISIC (descomprimir solo una vez)
+if not os.path.exists("saved_model"):
+    with zipfile.ZipFile("saved_model.zip", "r") as zip_ref:
+        zip_ref.extractall(".")
+
+# Cargar modelo ISIC con TensorFlow
+model_isic = tf.keras.models.load_model("saved_model")
+
+# Cargar modelos fastai
+model_malignancy = load_learner("modelo_malignancy.pkl")
+model_norm2000 = load_learner("modelo_norm2000.pkl")
+
+# Cargar modelo ViT
+from transformers import AutoImageProcessor, AutoModelForImageClassification
+feature_extractor = AutoImageProcessor.from_pretrained("nateraw/vit-skin-cancer")
+model_vit = AutoModelForImageClassification.from_pretrained("nateraw/vit-skin-cancer")
+
+# Clases y colores
+CLASSES = ['akiec', 'bcc', 'bkl', 'df', 'mel', 'nv', 'vasc']
+RISK_LEVELS = {
+    0: {"label": "akiec", "color": "#FF6F61", "weight": 0.9},
+    1: {"label": "bcc", "color": "#FF8C42", "weight": 0.7},
+    2: {"label": "bkl", "color": "#FFD166", "weight": 0.3},
+    3: {"label": "df", "color": "#06D6A0", "weight": 0.1},
+    4: {"label": "mel", "color": "#EF476F", "weight": 1.0},
+    5: {"label": "nv", "color": "#118AB2", "weight": 0.2},
+    6: {"label": "vasc", "color": "#073B4C", "weight": 0.4},
+}
+
+# Preprocesado para TensorFlow ISIC
+def preprocess_image_isic(pil_image):
+    image = pil_image.resize((224, 224))
+    array = np.array(image) / 255.0
+    return np.expand_dims(array, axis=0)
+
+# Función de análisis (como ya la tienes)
 def analizar_lesion_combined(img):
     try:
-        # Convertir imagen para Fastai
         img_fastai = PILImage.create(img)
-
-        # ViT prediction
         inputs = feature_extractor(img, return_tensors="pt")
         with torch.no_grad():
             outputs = model_vit(**inputs)
@@ -14,24 +60,19 @@ def analizar_lesion_combined(img):
         pred_class_vit = CLASSES[pred_idx_vit]
         confidence_vit = probs_vit[pred_idx_vit]
 
-        # Fast.ai models
         pred_fast_malignant, _, probs_fast_mal = model_malignancy.predict(img_fastai)
-        prob_malignant = float(probs_fast_mal[1])  # índice 1 = maligno
+        prob_malignant = float(probs_fast_mal[1])
         pred_fast_type, _, probs_fast_type = model_norm2000.predict(img_fastai)
 
-        # Modelo TensorFlow ISIC (usando TFSMLayer)
         x_isic = preprocess_image_isic(img)
         preds_isic_dict = model_isic(x_isic)
-
         print("🔍 Claves de salida de model_isic:", preds_isic_dict.keys())
-
         key = list(preds_isic_dict.keys())[0]
         preds_isic = preds_isic_dict[key].numpy()[0]
         pred_idx_isic = int(np.argmax(preds_isic))
         pred_class_isic = CLASSES[pred_idx_isic]
         confidence_isic = preds_isic[pred_idx_isic]
 
-        # Gráfico ViT
         colors_bars = [RISK_LEVELS[i]['color'] for i in range(7)]
         fig, ax = plt.subplots(figsize=(8, 3))
         ax.bar(CLASSES, probs_vit*100, color=colors_bars)
@@ -48,9 +89,7 @@ def analizar_lesion_combined(img):
         img_b64 = base64.b64encode(img_bytes).decode("utf-8")
         html_chart = f'<img src="data:image/png;base64,{img_b64}" style="max-width:100%"/>'
 
-        # Informe HTML con los 4 modelos
-        informe = f"""
-        <div style="font-family:sans-serif; max-width:800px; margin:auto">
+        informe = f"""<div style="font-family:sans-serif; max-width:800px; margin:auto">
         <h2>🧪 Diagnóstico por 4 modelos de IA</h2>
         <table style="border-collapse: collapse; width:100%; font-size:16px">
             <tr><th style="text-align:left">🔍 Modelo</th><th>Resultado</th><th>Confianza</th></tr>
@@ -58,10 +97,7 @@ def analizar_lesion_combined(img):
             <tr><td>🧬 Fast.ai (clasificación)</td><td><b>{pred_fast_type}</b></td><td>N/A</td></tr>
             <tr><td>⚠️ Fast.ai (malignidad)</td><td><b>{"Maligno" if prob_malignant > 0.5 else "Benigno"}</b></td><td>{prob_malignant:.1%}</td></tr>
             <tr><td>🔬 ISIC TensorFlow</td><td><b>{pred_class_isic}</b></td><td>{confidence_isic:.1%}</td></tr>
-        </table>
-        <br>
-        <b>🩺 Recomendación automática:</b><br> 
-        """
+        </table><br><b>🩺 Recomendación automática:</b><br>"""
 
         cancer_risk_score = sum(probs_vit[i] * RISK_LEVELS[i]['weight'] for i in range(7))
         if prob_malignant > 0.7 or cancer_risk_score > 0.6:
@@ -74,7 +110,6 @@ def analizar_lesion_combined(img):
             informe += "✅ <b>BAJO RIESGO</b> – Seguimiento de rutina (3-6 meses)"
 
         informe += "</div>"
-
         return informe, html_chart
 
     except Exception as e:
@@ -83,3 +118,17 @@ def analizar_lesion_combined(img):
         traceback.print_exc()
         return f"<b>Error interno:</b> {str(e)}", ""
 
+# INTERFAZ
+demo = gr.Interface(
+    fn=analizar_lesion_combined,
+    inputs=gr.Image(type="pil", label="Sube una imagen de la lesión"),
+    outputs=[gr.HTML(label="Informe combinado"), gr.HTML(label="Gráfico ViT")],
+    title="Detector de Lesiones Cutáneas (ViT + Fast.ai + ISIC TensorFlow)",
+    description="Comparación entre ViT transformer (HAM10000), dos modelos Fast.ai y el modelo ISIC TensorFlow.",
+    flagging_mode="never"
+)
+
+# LANZAMIENTO
+if __name__ == "__main__":
+    demo.launch()
+