Update app.py

app.py

@@ -22,8 +22,8 @@ from token_bucket import Limiter, MemoryStorage
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
 # Rate limiting configuration
-RATE_LIMIT = 10  # Requests per minute
-TOKEN_BUCKET_CAPACITY = RATE_LIMIT
+MAX_MODEL_CALLS_PER_MINUTE = 12  # Conservative buffer below 15 RPM
+RATE_LIMIT = MAX_MODEL_CALLS_PER_MINUTE
 TOKEN_BUCKET_REFILL_RATE = RATE_LIMIT / 60.0  # Tokens per second
 
 # Initialize global token bucket with MemoryStorage
@@ -56,7 +56,8 @@ async def submit_answers(session: aiohttp.ClientSession, submit_url: str,
             response.raise_for_status()
             return await response.json()
     except aiohttp.ClientResponseError as e:
-        print(f"Submission Failed: Server responded with status {e.status}. Detail: {e.message}")
+        print(f"Submission Failed: Server responded with status {e.status}. Detail: {e.message}"
+)
         return None
     except aiohttp.ClientError as e:
         print(f"Submission Failed: Network error - {e}")
@@ -65,40 +66,48 @@ async def submit_answers(session: aiohttp.ClientSession, submit_url: str,
         print(f"An unexpected error occurred during submission: {e}")
         return None
 
-async def process_question(agent, question_text: str, task_id: str,
- results_log: list):
+async def process_question(agent, question_text: str, task_id: str, results_log: list):
     """Process a single question with global rate limiting."""
     submitted_answer = None
-    try:
-        # Retry until a token is available
-        while not token_bucket.consume(1):
-            print(f"Rate limit reached for task {task_id}. Waiting to retry...")
-            await asyncio.sleep(60 / RATE_LIMIT)
-        submitted_answer = await agent(question_text)
-        results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
-        return {"task_id": task_id, "submitted_answer": submitted_answer}
-    except aiohttp.ClientResponseError as e:
-        if e.status == 429:
-            print(f"Rate limit hit for task {task_id}. Retrying after delay...")
-            await asyncio.sleep(60 / RATE_LIMIT)
-            while not token_bucket.consume(1):
-                await asyncio.sleep(60 / RATE_LIMIT)
-            try:
-                submitted_answer = await agent(question_text)
-                results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
-                return {"task_id": task_id, "submitted_answer": submitted_answer}
-            except Exception as retry_e:
-                submitted_answer = f"AGENT ERROR: {retry_e}"
+    max_retries = 3
+    retry_delay = 18  # Start with Gemini's recommended delay
+
+    for attempt in range(max_retries):
+        try:
+            if not token_bucket.consume(1):
+                print(f"Rate limit reached for task {task_id}. Waiting to retry...")
+                await asyncio.sleep(retry_delay)
+                continue
+
+            submitted_answer = await agent(question_text)
+            results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
+            return {"task_id": task_id, "submitted_answer": submitted_answer}
+            
+        except aiohttp.ClientResponseError as e:
+            if e.status == 429:
+                print(f"Rate limit hit for task {task_id}. Retrying after delay...")
+                retry_delay *= 2  # Exponential backoff
+                retry_delay += random.uniform(0, 5)  # Jitter
+                print(f"Retry #{attempt+1} in {retry_delay:.1f}s")
+                await asyncio.sleep(retry_delay)
+                while not token_bucket.consume(1):
+                    await asyncio.sleep(60 / RATE_LIMIT)
+                try:
+                    submitted_answer = await agent(question_text)
+                    results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
+                    return {"task_id": task_id, "submitted_answer": submitted_answer}
+                except Exception as retry_e:
+                    submitted_answer = f"AGENT ERROR: {retry_e}"
+                    results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
+                    return None
+            else:
+                submitted_answer = f"AGENT ERROR: {e}"
                 results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
                 return None
-        else:
+        except Exception as e:
             submitted_answer = f"AGENT ERROR: {e}"
             results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
             return None
-    except Exception as e:
-        submitted_answer = f"AGENT ERROR: {e}"
-        results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
-        return None
 
 async def run_and_submit_all(profile: gr.OAuthProfile | None):
     """
@@ -121,7 +130,7 @@ async def run_and_submit_all(profile: gr.OAuthProfile | None):
 
     # 1. Instantiate Agent
     try:
-        agent =MagAgent()
+        agent =MagAgent(rate_limiter=token_bucket)
     except Exception as e:
         print(f"Error instantiating agent: {e}")
         return f"Error initializing agent: {e}", None