fix: tenacity

requirements.txt

@@ -8,4 +8,5 @@ python-dotenv>=1.0.0
 agno==1.1.0
 pypdf>=3.11.1
 watchdog>=2.3.1
-sentencepiece>=0.1.99
+sentencepiece>=0.1.99
+tenacity>=8.2.2

services/model_handler.py

@@ -6,8 +6,11 @@ from agno.tools.arxiv import ArxivTools
 from agno.tools.pubmed import PubmedTools
 from agno.models.huggingface import HuggingFace
 import os
+import time
+from tenacity import retry, stop_after_attempt, wait_exponential
 
 MODEL_PATH = "google/flan-t5-small"
+
 class ModelHandler:
     def __init__(self):
         """Initialize the model handler"""
@@ -22,7 +25,13 @@ class ModelHandler:
     def _initialize_model(self):
         """Initialize model and tokenizer"""
         self.model, self.tokenizer = self._load_model()
-        base_model = HuggingFace(id=MODEL_PATH, api_key=os.getenv("hfToken"))
+        
+        try:
+            base_model = self._initialize_hf_model_with_retry()
+        except Exception as e:
+            logging.warning(f"Failed to initialize HuggingFace API model, falling back to local model: {str(e)}")
+            # Fallback to local model
+            base_model = self._initialize_local_model()
         
         self.translator = Agent(
             name="Translator",
@@ -42,7 +51,7 @@ class ModelHandler:
             instructions=[
                 "You need to understand the context of the question to provide the best answer based on your tools."
                 "Be precise and provide just enough information to be useful",
-                "You must cite the sources used in your answer."
+                "You must cite the sources used in your answer.",
                 "You must create an accessible summary.",
                 "The content must be for people without autism knowledge.",
                 "Focus in the main findings of the paper taking in consideration the question.",
@@ -104,6 +113,21 @@ class ModelHandler:
             logging.error(f"Error loading model: {str(e)}")
             return None, None
     
+    @retry(stop=stop_after_attempt(3), wait=wait_exponential(multiplier=1, min=4, max=10))
+    def _initialize_hf_model_with_retry(self):
+        """Initialize HuggingFace model with retry logic"""
+        api_key = os.getenv("hfToken")
+        if not api_key:
+            raise ValueError("HuggingFace API key not found in environment variables")
+        
+        return HuggingFace(id=MODEL_PATH, api_key=api_key)
+
+    def _initialize_local_model(self):
+        """Initialize local model as fallback"""
+        model = AutoModelForSeq2SeqLM.from_pretrained(MODEL_PATH)
+        tokenizer = AutoTokenizer.from_pretrained(MODEL_PATH)
+        return {"model": model, "tokenizer": tokenizer}
+    
     def generate_answer(self, query: str) -> str:
         try:
             translator = self.translator.run(query, stream=False)