Update app.py

app.py

@@ -11,7 +11,6 @@ from io import BytesIO
 from dotenv import load_dotenv
 import uvicorn
 
-# Carga de variables de entorno
 load_dotenv()
 
 API_KEY = os.getenv("API_KEY")
@@ -20,13 +19,11 @@ GOOGLE_APPLICATION_CREDENTIALS_JSON = os.getenv("GOOGLE_APPLICATION_CREDENTIALS_
 HF_API_TOKEN = os.getenv("HF_API_TOKEN")
 
 def validate_bucket_name(bucket_name):
-    """Valida que el nombre del bucket cumpla con las restricciones de Google Cloud."""
     if not re.match(r"^[a-z0-9][a-z0-9\-]*[a-z0-9]$", bucket_name):
         raise ValueError(f"Invalid bucket name '{bucket_name}'. Must start and end with a letter or number.")
     return bucket_name
 
 def validate_huggingface_repo_name(repo_name):
-    """Valida que el nombre del repositorio cumpla con las restricciones de Hugging Face."""
     if not re.match(r"^[a-zA-Z0-9_.-]+$", repo_name):
         raise ValueError(f"Invalid repository name '{repo_name}'. Must use alphanumeric characters, '-', '_', or '.'.")
     if repo_name.startswith(('-', '.')) or repo_name.endswith(('-', '.')) or '..' in repo_name:
@@ -35,15 +32,13 @@ def validate_huggingface_repo_name(repo_name):
         raise ValueError(f"Repository name '{repo_name}' exceeds max length of 96 characters.")
     return repo_name
 
-# Validar y configurar cliente de GCS
 try:
     GCS_BUCKET_NAME = validate_bucket_name(GCS_BUCKET_NAME)
     credentials_info = json.loads(GOOGLE_APPLICATION_CREDENTIALS_JSON)
     storage_client = storage.Client.from_service_account_info(credentials_info)
     bucket = storage_client.bucket(GCS_BUCKET_NAME)
 except (exceptions.DefaultCredentialsError, json.JSONDecodeError, KeyError, ValueError) as e:
-    print(f"Error al cargar credenciales o bucket: {e}")
-    exit(1)
+    raise RuntimeError(f"Error al cargar credenciales o bucket: {e}")
 
 app = FastAPI()
 
@@ -70,7 +65,6 @@ class GCSHandler:
         return BytesIO(blob.download_as_bytes())
 
 def download_model_from_huggingface(model_name):
-    """Descarga un modelo desde Hugging Face y lo sube a GCS."""
     model_name = validate_huggingface_repo_name(model_name)
     file_patterns = [
         "pytorch_model.bin",
@@ -89,7 +83,7 @@ def download_model_from_huggingface(model_name):
                 blob_name = f"{model_name}/{filename}"
                 bucket.blob(blob_name).upload_from_file(BytesIO(response.content))
         except Exception as e:
-            print(f"Error downloading {filename} from Hugging Face: {e}")
+            raise HTTPException(status_code=500, detail=f"Error downloading {filename} from Hugging Face: {e}")
 
 @app.post("/predict/")
 async def predict(request: DownloadModelRequest):