Hack machine language assembler

assembler.py

import re
from pathlib import Path

def main():
	dir = Path('.')
	for pi in dir.glob('**/*.asm'):
		po = Path(str(pi).replace('.asm', '.hack'))
		assemble(pi, po)

def assemble(pi, po):
	labels = {
		'SP': 0, 'LCL': 1, 'ARG': 2, 'THIS': 3, 'THAT': 4,
		'SCREEN': 16384, 'KBD': 24576,
	}
	for i in range(16):
		labels['R{}'.format(i)] = i
	VAR = 1024
	
	i = 0
	with pi.open() as fi:
		for l in fi:
			l = re.sub(r'//.*', '', l).strip()
			if not l: continue
			if l[0] == '(':
				labels[l[1:-1]] = i
				continue
			i += 1
	
	with po.open('w') as fo:
		with pi.open() as fi:
			i = 0
			for l in fi:
				l = re.sub(r'//.*', '', l).strip()
				if not l: continue
				if l[0] == '(': continue
				if l[0] == '@':
					sym = l[1:]
					try:
						sym = int(sym)
					except ValueError:
						if sym not in labels:
							labels[sym] = VAR
							VAR += 1
						sym = labels[sym]
					fo.write('{:016b}\n'.format(int(sym)))
				else:
					if ';' not in l:
						l += ';'
					if '=' not in l:
						l = '=' + l
					l = l.replace('=', ';')
					(dst, clc, jmp) = l.split(';')
					dst = DSTMAP[''.join(sorted(dst)).lower()]
					clc = clc.lower()
					jmp = JMPMAP[jmp.lower()]
					A = (1 if 'm' in clc else 0)
					clc = CLCMAP[clc.replace('m', 'a')]
					fo.write('111{}{}{}{}\n'.format(A, clc, dst, jmp))

JMPMAP = {
	'': '000', 'jmp': '111', 'jlt': '100',
	'jle': '110', 'jeq': '010', 'jge': '011',
	'jgt': '001', 'jne': '101',
}
DSTMAP = {
	'': '000', 'a': '100', 'd': '010',
	'm': '001', 'ad': '110', 'am': '101',
	'dm': '011', 'adm': '111',
}
CLCMAP = {
	'0': '101010',
	'1': '111111',
	'-1': '111010',
	'd': '001100',
	'a': '110000',
	'!d': '001101',
	'!a': '110001',
	'-d': '001111',
	'-a': '110011',
	'd+1': '011111',
	'a+1': '110111',
	'd-1': '001110',
	'a-1': '110010',
	'd+a': '000010',
	'd-a': '010011',
	'a-d': '000111',
	'd&a': '000000',
	'd|a': '010101',
}

if __name__ == '__main__':
	main()