TCPDashboard/database/repositories/raw_trade_repository.py

"""Repository for raw_trades table operations."""

from datetime import datetime, timedelta
from typing import Dict, Any, Optional, List

from sqlalchemy import desc, text

from ..models import RawTrade
from data.base_collector import MarketDataPoint
from .base_repository import BaseRepository, DatabaseOperationError


class RawTradeRepository(BaseRepository):
    """Repository for raw_trades table operations."""
    
    def insert_market_data_point(self, data_point: MarketDataPoint) -> bool:
        """
        Insert a market data point into raw_trades table using the ORM.
        """
        try:
            with self.get_session() as session:
                new_trade = RawTrade(
                    exchange=data_point.exchange,
                    symbol=data_point.symbol,
                    timestamp=data_point.timestamp,
                    data_type=data_point.data_type.value,
                    raw_data=data_point.data
                )
                session.add(new_trade)
                session.commit()
                
                self.log_debug(f"Stored raw {data_point.data_type.value} data for {data_point.symbol}")
                return True
                
        except Exception as e:
            self.log_error(f"Error storing raw data for {data_point.symbol}: {e}")
            raise DatabaseOperationError(f"Failed to store raw data: {e}")
    
    def insert_raw_websocket_data(self, 
                                 exchange: str,
                                 symbol: str, 
                                 data_type: str,
                                 raw_data: Dict[str, Any],
                                 timestamp: Optional[datetime] = None) -> bool:
        """
        Insert raw WebSocket data for debugging purposes using the ORM.
        """
        try:
            with self.get_session() as session:
                new_trade = RawTrade(
                    exchange=exchange,
                    symbol=symbol,
                    timestamp=timestamp or datetime.now(datetime.timezone.utc),
                    data_type=data_type,
                    raw_data=raw_data
                )
                session.add(new_trade)
                session.commit()
                
                self.log_debug(f"Stored raw WebSocket data: {data_type} for {symbol}")
                return True
                
        except Exception as e:
            self.log_error(f"Error storing raw WebSocket data for {symbol}: {e}")
            raise DatabaseOperationError(f"Failed to store raw WebSocket data: {e}")
    
    def get_raw_trades(self,
                      symbol: str,
                      data_type: str,
                      start_time: datetime,
                      end_time: datetime,
                      exchange: str = "okx",
                      limit: Optional[int] = None) -> List[Dict[str, Any]]:
        """
        Retrieve raw trades from the database using the ORM.
        """
        try:
            with self.get_session() as session:
                query = (
                    session.query(RawTrade)
                    .filter(
                        RawTrade.exchange == exchange,
                        RawTrade.symbol == symbol,
                        RawTrade.data_type == data_type,
                        RawTrade.timestamp >= start_time,
                        RawTrade.timestamp <= end_time
                    )
                    .order_by(RawTrade.timestamp.asc())
                )
                
                if limit:
                    query = query.limit(limit)

                results = query.all()

                trades = [
                    {
                        "id": r.id, "exchange": r.exchange, "symbol": r.symbol,
                        "timestamp": r.timestamp, "data_type": r.data_type,
                        "raw_data": r.raw_data, "created_at": r.created_at
                    } for r in results
                ]
                
                self.log_info(f"Retrieved {len(trades)} raw trades for {symbol} {data_type}")
                return trades
                
        except Exception as e:
            self.log_error(f"Error retrieving raw trades for {symbol}: {e}")
            raise DatabaseOperationError(f"Failed to retrieve raw trades: {e}")

    def cleanup_old_raw_data(self, days_to_keep: int = 7) -> int:
        """
        Clean up old raw data to prevent table bloat.
        
        Args:
            days_to_keep: Number of days to retain data.
            
        Returns:
            Number of records deleted.
        """
        try:
            cutoff_date = datetime.now(datetime.timezone.utc) - timedelta(days=days_to_keep)
            
            with self.get_session() as session:
                result = session.execute(
                    text("DELETE FROM raw_trades WHERE created_at < :cutoff_date"),
                    {"cutoff_date": cutoff_date}
                )
                deleted_count = result.rowcount
                session.commit()
                
                self.log_info(f"Cleaned up {deleted_count} old raw data records")
                return deleted_count
        except Exception as e:
            self.log_error(f"Failed to cleanup raw data: {e}")
            raise DatabaseOperationError(f"Failed to cleanup raw data: {e}")
    
    def get_raw_data_stats(self) -> Dict[str, Any]:
        """Get statistics about raw data storage."""
        try:
            with self.get_session() as session:
                result = session.execute(text("""
                    SELECT 
                        COUNT(*) as total_records,
                        COUNT(DISTINCT symbol) as unique_symbols,
                        COUNT(DISTINCT data_type) as data_types,
                        MIN(created_at) as oldest_record,
                        MAX(created_at) as newest_record,
                        pg_size_pretty(pg_total_relation_size('raw_trades')) as table_size
                    FROM raw_trades
                """)).fetchone()
                
                if result:
                    return dict(result._mapping)
                
                return {"status": "No data available"}
                
        except Exception as e:
            self.log_error(f"Failed to get raw data stats: {e}")
            raise DatabaseOperationError(f"Failed to get raw data stats: {e}")