Update app.py

app.py

@@ -4,6 +4,7 @@ from io import BytesIO
 from huggingface_hub import InferenceApi
 from gtts import gTTS
 import requests
+from requests.exceptions import ReadTimeout
 import tempfile
 import time
 import threading
@@ -62,7 +63,7 @@ def process_image(uploaded_file):
         st.error(f"Image processing error: {type(e).__name__}: {e}")
         st.stop()
 
-# —––––––– Helper: Generate Story with fallback —–––––––
+# —––––––– Helper: Generate Story with improved retry and timeout —–––––––
 def generate_story(prompt: str, caption: str) -> str:
     api_url = "https://api-inference.huggingface.co/models/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B"
     headers = {"Authorization": f"Bearer {hf_token}"}
@@ -79,9 +80,19 @@ def generate_story(prompt: str, caption: str) -> str:
     }
     retries = 0
     max_retries = 5
+    timeout = 60  # allow up to 60s for large model
     while True:
         try:
-            resp = requests.post(api_url, headers=headers, json=payload, timeout=30)
+            resp = requests.post(api_url, headers=headers, json=payload, timeout=timeout)
+        except ReadTimeout:
+            if retries < max_retries:
+                wait = 2 ** retries
+                st.info(f"Request timed out; retrying in {wait}s (attempt {retries+1}/{max_retries})")
+                time.sleep(wait)
+                retries += 1
+                continue
+            st.error("🚨 Story magic failed: request timed out after multiple attempts.")
+            st.stop()
         except Exception as e:
             st.error(f"🚨 Story magic failed: {type(e).__name__}: {e}")
             st.stop()
@@ -100,7 +111,7 @@ def generate_story(prompt: str, caption: str) -> str:
 
         # Model loading (cold start)
         if resp.status_code == 503 and retries < max_retries:
-            wait = int(resp.json().get("estimated_time", 5)) if resp.headers.get('Content-Type','').startswith('application/json') else 5 * (2 ** retries)
+            wait = int(resp.json().get("estimated_time", 5))
             st.info(f"Model loading; retrying in {wait}s (attempt {retries+1}/{max_retries})")
             time.sleep(wait)
             retries += 1
@@ -108,12 +119,12 @@ def generate_story(prompt: str, caption: str) -> str:
 
         # Server-side generation error
         if resp.status_code in (424, 500, 502) and retries < max_retries:
-            st.info(f"Server error {resp.status_code}; retrying (attempt {retries+1}/{max_retries})")
-            time.sleep(2 ** retries)
+            wait = 2 ** retries
+            st.info(f"Server error {resp.status_code}; retrying in {wait}s (attempt {retries+1}/{max_retries})")
+            time.sleep(wait)
             retries += 1
             continue
         if resp.status_code in (424, 500, 502):
-            # Fallback story using the caption, ensuring ~70 words
             return (f"One day, {caption} woke up under a bright sky and decided to explore the garden. "
                     "It met a friendly ladybug and together they played hide-and-seek among the flowers. "
                     "At sunset, {caption} curled up by a daisy, purring happily as it dreamed of new adventures.")
@@ -177,3 +188,4 @@ if uploaded:
 
 # Footer
 st.markdown("---\n*Made with ❤️ by your friendly story wizard*")
+