refactor: enhance Optuna backtesting framework, optimize orderbook filtering, and update database management utilities.

scratch/delete_old_data_pymysql.py

@@ -0,0 +1,68 @@
+import sys
+import time
+import pymysql
+
+conn = pymysql.connect(
+    host='192.168.0.141',
+    port=3306,
+    user='jae',
+    password='1234',
+    database='kis_quant_db',
+    cursorclass=pymysql.cursors.DictCursor,
+    autocommit=True
+)
+
+try:
+    with conn.cursor() as cur:
+        cur.execute("SHOW TABLES")
+        tables = [list(r.values())[0] for r in cur.fetchall()]
+        
+    target_prefixes = (
+        'ws_ticks', 'ws_orderbook', 'ws_candles', 
+        'ls_ws_ticks', 'ls_ws_orderbook', 'ls_ws_candles',
+        'ws_price_validation'
+    )
+    
+    for table in tables:
+        if not table.startswith(target_prefixes):
+            continue
+            
+        with conn.cursor() as cur:
+            cur.execute(f"SHOW COLUMNS FROM {table}")
+            cols = [r["Field"] for r in cur.fetchall()]
+            
+        time_col = None
+        for candidate in ('tick_time', 'ob_time', 'candle_time', 'timestamp', 'recv_ts', 'datetime'):
+            if candidate in cols:
+                time_col = candidate
+                break
+                
+        if time_col:
+            cutoff = '20260808'
+            
+            with conn.cursor() as cur:
+                count_q = f"SELECT COUNT(*) as cnt FROM {table} WHERE {time_col} < %s"
+                cur.execute(count_q, (cutoff,))
+                count = cur.fetchone()['cnt']
+                
+            if count > 0:
+                print(f"Deleting {count} rows from {table} (using column {time_col})")
+                deleted = 0
+                while True:
+                    with conn.cursor() as cur:
+                        del_q = f"DELETE FROM {table} WHERE {time_col} < %s LIMIT 100000"
+                        cur.execute(del_q, (cutoff,))
+                        rows_affected = cur.rowcount
+                        
+                    if rows_affected <= 0:
+                        break
+                    deleted += rows_affected
+                    print(f"  ... deleted {deleted}/{count} rows in {table}")
+                    time.sleep(0.01)
+                print(f"Finished deleting from {table}.")
+            else:
+                print(f"No rows to delete in {table} before {cutoff}")
+        else:
+            print(f"Warning: No time column found for {table}, columns: {cols}")
+finally:
+    conn.close()