refine

msr.py

@@ -209,7 +209,6 @@ def get_duckdb_connection():
     conn.execute("SET memory_limit = '400GB';")  # Utilize available RAM (709GB available)
     conn.execute("SET preserve_insertion_order = false;")  # Better parallelization
     conn.execute("SET enable_object_cache = true;")  # Cache objects for reuse
-    conn.execute("SET force_parallelism = true;")  # Force parallel execution where possible
     conn.execute("SET temp_directory = '/tmp/duckdb_temp';")  # Use fast temp storage if needed
 
     return conn
@@ -1047,7 +1046,6 @@ def mine_all_agents():
     print(f"{'='*80}")
     print(f"Starting review metadata mining for {len(identifiers)} agents")
     print(f"Time frame: Last {LEADERBOARD_TIME_FRAME_DAYS} days")
-    print(f"Data source: DuckDB + Local GHArchive (SINGLE QUERY)")
     print(f"{'='*80}")
 
     # Initialize DuckDB connection

requirements.txt

@@ -1,6 +1,6 @@
 APScheduler
 backoff
-duckdb
+duckdb[all]
 gradio
 gradio_leaderboard
 huggingface_hub