Truncate file names

I noticed that if you do a very long query the file names that get generated can be larger than the host permits. So I added some truncation that should work pretty universally to limit the file name to under <250 characters total including the important naming convention of the prefixes.

backend/server/server_utils.py

@@ -99,7 +99,20 @@ async def research(self) -> dict:
         }
 
 def sanitize_filename(filename: str) -> str:
-    return re.sub(r"[^\w\s-]", "", filename).strip()
+    # Split into components
+    prefix, timestamp, *task_parts = filename.split('_')
+    task = '_'.join(task_parts)
+
+    # Calculate max length for task portion
+    # 255 - len("outputs/") - len("task_") - len(timestamp) - len("_.json") - safety_margin
+    max_task_length = 255 - 8 - 5 - 10 - 6 - 10  # ~216 chars for task
+
+    # Truncate task if needed
+    truncated_task = task[:max_task_length] if len(task) > max_task_length else task
+
+    # Reassemble and clean the filename
+    sanitized = f"{prefix}_{timestamp}_{truncated_task}"
+    return re.sub(r"[^\w\s-]", "", sanitized).strip()
 
 
 async def handle_start_command(websocket, data: str, manager):