Update app.py

app.py

@@ -12,6 +12,10 @@ from langchain_community.vectorstores import FAISS
 from langchain.prompts import PromptTemplate
 import os
 from dotenv import load_dotenv
+from requests.adapters import HTTPAdapter
+from requests.packages.urllib3.util.retry import Retry
+import requests
+import time
 
 # Load environment variables
 load_dotenv()
@@ -61,25 +65,37 @@ class DarijaTherapist:
                 device=self.device
             )
             
-            # LLM setup
+            # Configure retry strategy
+            retry_strategy = Retry(
+                total=3,  # number of retries
+                backoff_factor=1,  # wait 1, 2, 4 seconds between retries
+                status_forcelist=[429, 500, 502, 503, 504]  # HTTP status codes to retry on
+            )
+            
+            # Create session with retry strategy
+            session = requests.Session()
+            session.mount("https://", HTTPAdapter(max_retries=retry_strategy))
+            
+            # LLM setup with increased timeout and retry session
             self.llm = HuggingFaceEndpoint(
-                endpoint_url="https://api-inference.huggingface.co/models/MBZUAI-Paris/Atlas-Chat-2B",
+                endpoint_url="https://api-inference.huggingface.co/models/MBZUAI-Paris/Atlas-Chat-9B",
                 task="text-generation",
                 temperature=0.7,
                 do_sample=True,
                 return_full_text=False,
                 model_kwargs={
                     "max_length": 512,
+                    "timeout": 300  # increase timeout to 5 minutes
                 },
-                huggingfacehub_api_token=os.getenv("HUGGINGFACE_API_TOKEN")
+                huggingfacehub_api_token=os.getenv("HUGGINGFACE_API_TOKEN"),
+                client=session  # use configured session
             )
             
-            # Embeddings setup
+            # Rest of the setup remains the same
             self.embeddings = HuggingFaceBgeEmbeddings(
                 model_name="BAAI/bge-large-en"
             )
             
-            # Vector store setup
             self.vectorstore = FAISS.from_texts(
                 ["Initial therapeutic context"], 
                 self.embeddings
@@ -145,15 +161,23 @@ class DarijaTherapist:
             st.session_state.audio_buffer = []
 
     def get_ai_response(self, user_input):
-        try:
-            response = self.conversation_chain({
-                "question": user_input,
-                "chat_history": self.memory.chat_memory.messages
-            })
-            return response['answer']
-        except Exception as e:
-            st.error(f"Error getting AI response: {str(e)}")
-            return "عذراً، كاين شي مشكل. حاول مرة أخرى."
+        max_retries = 3
+        for attempt in range(max_retries):
+            try:
+                response = self.conversation_chain({
+                    "question": user_input,
+                    "chat_history": self.memory.chat_memory.messages
+                })
+                return response['answer']
+            except requests.exceptions.ReadTimeout:
+                if attempt < max_retries - 1:
+                    st.warning(f"Attempt {attempt + 1} timed out, retrying...")
+                    time.sleep(2 ** attempt)  # exponential backoff
+                    continue
+                return "عذراً، الخادم بطيء حالياً. حاول مرة أخرى."
+            except Exception as e:
+                st.error(f"Error getting AI response: {str(e)}")
+                return "عذراً، كاين شي مشكل. حاول مرة أخرى."
 
     def run(self):
         st.set_page_config(page_title="Darija AI Therapist", page_icon="🧠")