Refactor API client initialization and add comprehensive API configuration settings

app.py

@@ -235,7 +235,10 @@ def initialize_client(model_id=None):
     
     client = InferenceClient(
         model_id,
-        token=HF_TOKEN
+        token=API_CONFIG["token"],
+        endpoint=API_CONFIG["inference_endpoint"],
+        headers=API_CONFIG["headers"],
+        timeout=API_CONFIG["timeout"]
     )
     return client
 

config/settings.py

@@ -5,6 +5,18 @@ HF_TOKEN = os.getenv("HUGGINGFACE_TOKEN")
 if not HF_TOKEN:
     raise ValueError("HUGGINGFACE_TOKEN not found in environment variables")
 
+# API Configuration
+API_CONFIG = {
+    "inference_endpoint": os.getenv("HF_INFERENCE_ENDPOINT", "https://api-inference.huggingface.co"),
+    "token": HF_TOKEN,
+    "is_paid_tier": True,  # или False в зависимости от вашего плана
+    "timeout": 30,
+    "headers": {
+        "X-Use-Cache": "false",
+        "Content-Type": "application/json"
+    }
+}
+
 # Dataset configuration
 DATASET_ID = "Rulga/status-law-knowledge-base"
 CHAT_HISTORY_PATH = "chat_history"
@@ -111,6 +123,10 @@ MODELS = {
     }
 }
 
+# Update MODELS configuration
+for model in MODELS.values():
+    model["endpoint"] = API_CONFIG["inference_endpoint"]
+
 # Default model
 DEFAULT_MODEL = "llama-7b"  # Changed from "zephyr-7b" to "llama-7b"
 ACTIVE_MODEL = MODELS[DEFAULT_MODEL]

test_endpoint.py

@@ -0,0 +1,32 @@
+from huggingface_hub import HfApi, InferenceClient
+import os
+
+# Ваш текущий токен
+token = os.getenv("HUGGINGFACE_TOKEN")
+
+# Проверка типа доступа
+api = HfApi(token=token)
+try:
+    # Проверяем информацию об аккаунте
+    user_info = api.whoami()
+    print("Account type:", user_info.get("type"))
+    print("Plan:", user_info.get("plan"))
+    
+    # Проверяем доступные эндпоинты
+    endpoints = api.list_endpoints()
+    print("\nAvailable endpoints:")
+    for endpoint in endpoints:
+        print(f"- {endpoint.name}: {endpoint.url}")
+except Exception as e:
+    print(f"Error checking endpoints: {e}")
+
+# Проверяем текущий клиент
+client = InferenceClient(
+    "HuggingFaceH4/zephyr-7b-beta",  # или ваша текущая модель
+    token=token
+)
+
+# Проверяем тип подключения
+print("\nClient information:")
+print("API Base URL:", client.api_url)
+print("Headers:", client.headers)