Best way to match a C++ struct in python?

Question

I have an API endpoint that takes a binary struct on a server written in C++. The endpoint must accept hundreds of these structs per second so using JSON or something similar has too much overhead.

The struct looks something like this:

struct TX {
    int blockId;
    char signature[64];
    char signingKey[32];
    time_t timestamp;
    char nonce[8];
    char to[25];
    char from[25];
    long amount;
    long fee;
    bool isTransactionFee;
};

I am writing a client in Python that must submit data in the above format. When I do sizeof on the struct in C++ it gives me 200 bytes, the sum of the individual fields sizeof's is somewhere closer to 182 bytes.

I am not sure where to start in terms of being able to serialize the structs in python accurately -- Do I assume that the padding to the struct is always added to the end? How do I know there isn't padding in between one of the two fields?

I wrote something along these lines to get started and plan to copy the data into these bytearrays:

    txBinary = bytearray([0]*200)
    blockId = bytearray([0]*4)
    signature = bytearray([0]*64)
    signingKey = bytearray([0]*32)
    timestamp = bytearray([0]*8)
    nonce = bytearray([0]*8)
    toWallet = bytearray([0]*25)
    fromWallet = bytearray([0]*25)
    amount = bytearray([0]*8)
    fee = bytearray([0]*8)
    isFee = bytearray([0])

Is this is a sane approach? Is there some other easier way to serialize a datastructure like this in Python?

Answer 1

So after a lot of finagling, I decided the most robust way to do this is to use exec from within Python to call a C++ cli app that takes in the transaction parameters and outputs a base64 encoded version of the struct.

The most difficult challenge was that I needed to compute a cryptographic signature over the bytes in the transaction and getting the C++ and python versions of the cryptography libraries to work in unison became a pain, this led me to just use the C++ codebase and call it from within Python.