[Python] How to Correctly Decode a Jupiter Swap Event on Solana?

I am attempting to decode a Jupiter swap event on the Solana blockchain to extract token amounts and mint addresses from the transaction data. The expected output should look like this:

{
"amm": "whirLbMiicVdio4qvUfM5KAg6Ct8VwpYzGff3uctyCc",
"inputMint": "So11111111111111111111111111111111111111112",
"inputAmount": "14829998",
"outputMint": "Hjw6bEcHtbHGpQr8onG3izfJY5DJiWdt7uk2BfdSpump",
"outputAmount": "127849833"
}


Here is an exampleTransaction for reference.

My Approach

I am using construct to define the expected data structure and parse the event:

from construct import Int64ul, Bytes
from construct import Struct as cStruct
import base58

PUBLIC_KEY_LAYOUT = Bytes(32)

JUP_SWAP = cStruct(
"amm" / PUBLIC_KEY_LAYOUT,
"inputMint" / PUBLIC_KEY_LAYOUT,
"inputAmount" / Int64ul,
"outputMint" / PUBLIC_KEY_LAYOUT,
"outputAmount" / Int64ul,
)

# Raw transaction data (copied from Solscan in "raw" mode)
hex_data = 'e445a52e51cb9a1d40c6cde8260871e20e03685f8e909053e458121c66f5a76aedc7706aa11c82f8aa952a8f2b7879a9069b8857feab8184fb687f634618c035dac439dc1aeb3b5598a0f00000000001ae49e20000000000f8ba8d41600d59ff5ede65103bfc142e5bc863e0e17ce50cc1155c425f02395f69d59e0700000000'

# Convert hex to bytes and remove the first 8 bytes
bytes_event = bytes.fromhex(hex_data)[8:]

# Decode the event
decoded_event = JUP_SWAP.parse(bytes_event)
amm = base58.b58encode(decoded_event.amm).decode()
input_mint = base58.b58encode(decoded_event.inputMint).decode()
input_amount = decoded_event.inputAmount
output_mint = base58.b58encode(decoded_event.outputMint).decode()
output_amount = decoded_event.outputAmount

print(f"amm: {amm}")
print(f"inputMint: {input_mint}")
print(f"inputAmount: {input_amount}")
print(f"outputMint: {output_mint}")
print(f"outputAmount: {output_amount}")


Issue

The output does not match the expected values:

amm: 5Mrv98ERNPgDKHDFz8AjcebvXyf8dq9kp1yY1RAHSCJB
inputMint: CUtETTsHosks4AUNtZuBZqb4Gn9k7JetykayaCyDcFq6
inputAmount: 72057594053697688
outputMint: CjMHLni4ZdtAi91Am1Vx63jCYL2bpEArkqzjzsBuaqMV
outputAmount: 6861518115434141121


I’ve tried multiple other decoding methods, including:

Attempt 1: Using PublicKey and struct

amm = PublicKey(buffer[0:32]).__str__()
input_mint = PublicKey(buffer[32:64]).__str__()
input_amount = struct.unpack("<Q", buffer[64:72])[0] # Little-endian u64
output_mint = PublicKey(buffer[72:104]).__str__()
output_amount = struct.unpack("<Q", buffer[104:112])[0] # Little-endian u64


Attempt 2: Using base58 encoding

amm = base58.b58encode(buffer[0:32]).decode()
inputMint = base58.b58encode(buffer[32:64]).decode()
inputAmount = int.from_bytes(buffer[64:72],'little')
outputMint = base58.b58encode(buffer[72:104]).decode()
outputAmount = int.from_bytes(buffer[104:112],'little')


Attempt 3: Using struct.unpack

unpacked = struct.unpack("=32s32sQ32sQ", buffer)
amm = base58.b58encode(unpacked[0]).decode()
inputMint = base58.b58encode(unpacked[2]).decode()
inputAmount = unpacked[2]
outputMint = base58.b58encode(unpacked[3]).decode()
outputAmount = unpacked[4]


Regardless of the method, the results remain incorrect.

Question

• What is the correct way to decode a Jupiter swap event on Solana? Am
• I missing a step in parsing the transaction data?
• Could there be an issue with endianness or byte alignment?

I referenced the expected event structure from this IDL, but my results do not align.

Any guidance or corrections would be greatly appreciated. Thanks in advance!

Continue reading...