summarize_stream.py

import struct
import sys

def summarize_stream(file_path):
    """
    Parses and summarizes the content of a binary .bin file.
    """
    print(f"Summarizing file: {file_path}")
    try:
        with open(file_path, 'rb') as f:
            count = 0
            checkpoint_count = 100000
            while True:
                count += 1
                if count % checkpoint_count == 0:
                    # Save to checkpoint ckpt_{count}.bin
                    
                    print(f"Processed {count} messages...")

                header = f.read(8)
                if not header:
                    break
                
                if len(header) < 8:
                    print(f"Incomplete header found: {header.hex()}")
                    break

                seq_no, msg_size = struct.unpack('<II', header)
                
                msg_data = f.read(msg_size)
                if len(msg_data) < msg_size:
                    print(f"Incomplete message data for seq_no {seq_no}. Expected {msg_size}, got {len(msg_data)}")
                    break

                msg_type = chr(msg_data[0])
                
                print(f"SeqNo: {seq_no}, MsgSize: {msg_size}, MsgType: {msg_type}", end='')

                if msg_type == 'A' or msg_type == 'U':
                    # ADD or UPDATE
                    symbol = msg_data[1:4].decode('ascii').strip()
                    order_id = struct.unpack('<Q', msg_data[4:12])[0]
                    side = chr(msg_data[12])
                    volume = struct.unpack('<Q', msg_data[16:24])[0]
                    price = struct.unpack('<i', msg_data[24:28])[0]
                    print(f", Symbol: {symbol}, OrderId: {order_id}, Side: {side}, Volume: {volume}, Price: {price}")

                elif msg_type == 'D':
                    # DELETE
                    symbol = msg_data[1:4].decode('ascii').strip()
                    order_id = struct.unpack('<Q', msg_data[4:12])[0]
                    side = chr(msg_data[12])
                    print(f", Symbol: {symbol}, OrderId: {order_id}, Side: {side}")

                elif msg_type == 'E':
                    # EXECUTE
                    symbol = msg_data[1:4].decode('ascii').strip()
                    order_id = struct.unpack('<Q', msg_data[4:12])[0]
                    side = chr(msg_data[12])
                    traded_qty = struct.unpack('<Q', msg_data[16:24])[0]
                    print(f", Symbol: {symbol}, OrderId: {order_id}, Side: {side}, TradedQty: {traded_qty}")
                else:
                    print(f", Unknown message type: {msg_type}")

    except FileNotFoundError:
        print(f"Error: File not found at {file_path}")
    except Exception as e:
        print(f"An error occurred: {e}")
    print("-" * 20)

if __name__ == "__main__":
    if len(sys.argv) > 1:
        for file_path in sys.argv[1:]:
            summarize_stream(file_path)
    else:
        print("Usage: python summarize_stream.py <file1.bin> [<file2.bin> ...]")