Update app.py

app.py

@@ -1,11 +1,12 @@
 import gradio as gr
 import requests
 import json
+import time
 from faster_whisper import WhisperModel
 
 OLLAMA_URL = "http://localhost:11434"
 
-# UPDATED: Pointing to the working bartowski GGUF repo
+# Pointing to the working bartowski GGUF repo
 MODELS = {
     "Qwen2.5-Coder 1.5B (Fastest)": "qwen2.5-coder:1.5b",
     "Qwen2.5-Coder 3B (Fast)": "qwen2.5-coder:3b",
@@ -24,12 +25,19 @@ def check_ollama():
     except:
         return False
 
-# NEW FUNCTION: Auto-download the model if missing
+# NEW FUNCTION: Robust startup that waits for Ollama
 def ensure_model(model_name):
-    if not check_ollama():
-        print("❌ Ollama not running, skipping model download.")
-        return
+    print(f"⏳ Waiting for Ollama to start...")
     
+    # Wait up to 30 seconds for Ollama to be ready
+    retries = 0
+    while not check_ollama():
+        time.sleep(2)
+        retries += 1
+        if retries > 15:
+            print("❌ Ollama failed to start in time.")
+            return
+
     print(f"🔎 Checking for model: {model_name}")
     try:
         # Check if model is already loaded
@@ -39,7 +47,7 @@ def ensure_model(model_name):
             return
         
         # If not, pull it
-        print(f"📥 Downloading {model_name}... (This may take a few minutes)")
+        print(f"📥 Downloading {model_name}... (This WILL take time for 30B)")
         with requests.post(f"{OLLAMA_URL}/api/pull", json={"name": model_name}, stream=True) as r:
             for line in r.iter_lines():
                 pass