SCS assembler

asm.py

#!/usr/bin/env python3
import re
import sys
import pprint

REG_BIT = {
    "R0": 0 << 4,
    "R1": 1 << 4
}

def imm(imm):
    return int(imm[1:], 0) & 0b1111

def encode_ADD(*_):
    return 0b00100000

def encode_MOV(arg1, arg2):
    if arg2[0] == "#":
        return 0b10000000 | REG_BIT[arg1] | imm(arg2)
    else:
        return 0b10100000 | REG_BIT[arg1]

def encode_LDR(arg1, arg2):
    return 0b11000000 | REG_BIT[arg1] | imm(arg2)

def encode_STR(arg1, arg2):
    return 0b11100000 | REG_BIT[arg1] | imm(arg2)

def encode_JMP(arg1, *_):
    return 0b00000000 | imm(arg1)

ENCODERS = {
    "ADD": encode_ADD,
    "MOV": encode_MOV,
    "LDR": encode_LDR,
    "STR": encode_STR,
    "JMP": encode_JMP
}

def encode(statements):
    return [
        ENCODERS[instr](arg1, arg2)
        for instr, arg1, arg2 in statements
    ]

STMT_PAT = re.compile(r"(\w+)(?: +([\w#]+)(?: *, *([\w#]+))?)?")

def parse_file(file):
    return [
        [m.group(i) for i in (1, 2, 3)]
        for m in [
            STMT_PAT.match(line)
            for line in file.readlines()
        ]
        if m is not None
    ]

JMP_ZERO = encode_JMP("#0")

def format_vector(program):
    filled_program = program + [JMP_ZERO] * (16 - len(program))
    return ",\n".join(format(stmt, "02X") for stmt in filled_program) + ";"

def main():
    path = sys.argv[1] if len(sys.argv) > 1 else "prog.txt"

    with open(path) as file:
        statements = parse_file(file)
        program = encode(statements)

        pprint.pprint(statements)
        print("-" * 20)
        print(format_vector(program))

if __name__ == "__main__":
    main()