2019-12-21 - Springdroid Adventure
(original .ipynb)
Day 21 puzzle input is an IntCode program (mine is here) which represents a networked computer as part of a bigger network. Part 1 involves connecting a number of these network computers and finding the first
opcode_add = 1
opcode_mul = 2
opcode_read = 3
opcode_write = 4
opcode_jump_true = 5
opcode_jump_false = 6
opcode_lt = 7
opcode_eq = 8
opcode_rebase = 9
opcode_terminate = 99
mode_position = 0
mode_immediate = 1
mode_relative = 2
class IntCodeCpu:
def __init__(self, memory_image):
self.memory = [ x for x in memory_image ] # copy memory image, in case it's reused
self.stalled = True
self.input_buffer = None
self.output_buffer = None
self.pc = 0
self.initialise_opcodes()
self.offset = 0
self.done = False
def start(self, input_buffer, output_buffer, noun=None, verb=None):
self.input_buffer = input_buffer
self.output_buffer = output_buffer
if noun:
self.memory[1] = noun
if verb:
self.memory[2] = verb
return self.run()
def run(self):
instr = self.memory[self.pc]
self.stalled = False
while int(instr) != opcode_terminate and not self.stalled:
(op, modes) = self.decode_instr(instr)
self.pc = op(modes)
instr = self.memory[self.pc]
return self.memory[0]
#-HELPERS-----------------------------
def try_pop_mode(self, modes):
if len(modes) == 0:
return 0
return modes.pop()
def resize_memory(self, target_addr):
self.memory += ([0] * (1 + target_addr - len(self.memory)))
#-DECODE-INSTRUCTIONS-----------------
def initialise_opcodes(self):
self.opcodes = {
opcode_add: self.op_add,
opcode_mul: self.op_mul,
opcode_read: self.op_read,
opcode_write: self.op_write,
opcode_jump_true: self.op_jump_true,
opcode_jump_false: self.op_jump_false,
opcode_lt: self.op_lt,
opcode_eq: self.op_eq,
opcode_rebase: self.op_rebase
}
def decode_instr(self, instr):
instr = str(instr)
opcode = self.decode_op(instr)
modes = self.decode_modes(instr)
if not (opcode in self.opcodes):
raise Exception(f"Invalid opcode {opcode}")
return (self.opcodes[opcode], modes)
def decode_op(self, instr):
if len(instr) > 2:
return int(instr[-2:])
return int(instr)
def decode_modes(self, instr):
if len(instr) > 2:
return [ int(d) for d in instr[:-2]]
return []
#-MICRO-OPS---------------------------
def uop_read(self, value, mode):
if mode == mode_position:
if value >= len(self.memory):
self.resize_memory(value)
return int(self.memory[value])
elif mode == mode_relative:
if self.offset + value >= len(self.memory):
self.resize_memory(self.offset + value)
return int(self.memory[self.offset + value])
elif mode == mode_immediate:
return int(value)
else:
raise Exception("UNKNOWN MODE")
def uop_write(self, dst, value, mode):
if mode == mode_position:
if dst >= len(self.memory):
self.resize_memory(dst)
self.memory[dst] = value
elif mode == mode_relative:
if self.offset + dst >= len(self.memory):
self.resize_memory(self.offset + dst)
self.memory[self.offset + dst] = value
elif mode == mode_immediate:
raise Exception(f"cannot write {value} to literal {dst}")
def uop_cond_jump(self, modes, cond):
param_mode = self.try_pop_mode(modes)
param_raw = int(self.memory[self.pc + 1])
param = self.uop_read(param_raw, param_mode)
dest_mode = self.try_pop_mode(modes)
dest_raw = int(self.memory[self.pc + 2])
dest = self.uop_read(dest_raw, dest_mode)
if cond(param):
return dest
return self.pc + 3
def uop_cmp(self, modes, cmp):
param0_mode = self.try_pop_mode(modes)
param0_raw = int(self.memory[self.pc + 1])
param0 = self.uop_read(param0_raw, param0_mode)
param1_mode = self.try_pop_mode(modes)
param1_raw = int(self.memory[self.pc + 2])
param1 = self.uop_read(param1_raw, param1_mode)
dest_mode = self.try_pop_mode(modes)
dest = int(self.memory[self.pc + 3])
if cmp(param0, param1):
self.uop_write(dest, 1, dest_mode)
else:
self.uop_write(dest, 0, dest_mode)
return self.pc + 4
#-OPCODES-----------------------------
def op_add(self, modes):
arg0_mode = self.try_pop_mode(modes)
arg1_mode = self.try_pop_mode(modes)
dest_mode = self.try_pop_mode(modes)
arg0_raw = int(self.memory[self.pc + 1])
arg1_raw = int(self.memory[self.pc + 2])
dest = int(self.memory[self.pc + 3])
arg0 = self.uop_read(arg0_raw, arg0_mode)
arg1 = self.uop_read(arg1_raw, arg1_mode)
self.uop_write(dest, str(int(arg0) + int(arg1)), dest_mode)
return self.pc + 4
def op_mul(self, modes):
arg0_mode = self.try_pop_mode(modes)
arg1_mode = self.try_pop_mode(modes)
dest_mode = self.try_pop_mode(modes)
arg0_raw = int(self.memory[self.pc + 1])
arg1_raw = int(self.memory[self.pc + 2])
dest = int(self.memory[self.pc + 3])
arg0 = self.uop_read(arg0_raw, arg0_mode)
arg1 = self.uop_read(arg1_raw, arg1_mode)
self.uop_write(dest, str(int(arg0) * int(arg1)), dest_mode)
return self.pc + 4
def op_read(self, modes):
dest_mode = self.try_pop_mode(modes)
dest = int(self.memory[self.pc + 1])
# if the input buffer is empty, we should "stall" and
# resume later
if not self.input_buffer:
self.stalled = True
return self.pc
val = self.input_buffer.pop()
self.uop_write(dest, str(val), dest_mode)
return self.pc + 2
def op_write(self, modes):
src_mode = self.try_pop_mode(modes)
src_raw = int(self.memory[self.pc + 1])
src = self.uop_read(src_raw, src_mode)
self.output_buffer.append(src)
return self.pc + 2
def op_jump_true(self, modes):
return self.uop_cond_jump(modes, lambda x: x != 0)
def op_jump_false(self, modes):
return self.uop_cond_jump(modes, lambda x: x == 0)
def op_lt(self, modes):
return self.uop_cmp(modes, lambda x, y: x < y)
def op_eq(self, modes):
return self.uop_cmp(modes, lambda x, y: x == y)
def op_rebase(self, modes):
param_mode = self.try_pop_mode(modes)
param_raw = int(self.memory[self.pc + 1])
param = self.uop_read(param_raw, param_mode)
self.offset += param
return self.pc + 2
raw_code = open("puzzle_input/day21.txt", "r").read()
code = raw_code.split(",")
def preprocess_ascii_input(instructions):
int_instructions = [ ord(c) for c in instructions ]
return int_instructions + [ ord("\n") ]
def skip_line(line):
return len(line) == 0 or line[0] == "#"
def run_script(script):
output_buffer = []
cpu = IntCodeCpu(code)
script_lines = [ preprocess_ascii_input(script_line) for script_line in script.split("\n") if not skip_line(script_line.strip()) ]
from pprint import pprint
output = []
for i, script_line in enumerate(script_lines):
print(i, script_line)
cpu.start(script_line[::-1], output_buffer)
maybe_return_value = output_buffer[-1]
if (maybe_return_value >= 0 and maybe_return_value < 256):
# the output is printable ascii meaning that we
# have hit a snag - we should print the output
# to see the last dying moments of our dear
# springbot
output = "".join([ chr(c) for c in output_buffer ])
print(output)
else:
print("retval = ", maybe_return_value)
# part 1 # reached via some very simple heuristics - if there is # an gap upcoming just before jump target (D) we should jump: # # @ # #??.# # ABCD # # and if there's a gap immediately in front of us, we should jump # # @ # #.??? # ABCD # # otherwise, we'll walk script_part_1 = """ # if there's a gap *just* before jump target, jump NOT C T AND D T OR T J # if there's a gap right in front of us, jump NOT A T OR T J # fingers crossed WALK""" run_script(script_part_1)
0 [78, 79, 84, 32, 67, 32, 84, 10] 1 [65, 78, 68, 32, 68, 32, 84, 10] 2 [79, 82, 32, 84, 32, 74, 10] 3 [78, 79, 84, 32, 65, 32, 84, 10] 4 [79, 82, 32, 84, 32, 74, 10] 5 [87, 65, 76, 75, 10] retval = 19354392
# part 2 # # more tricky, this time we need to be more careful about our # fallback of "otherwise, we'll walk" # # ......@.......... # #####.##.##..#### # ABCDEFGH # # here the existing program believes we can go ahead - but we # end up being forced to jump. we need to jump before this but # how do we decide this? a naive guess would be "if there's yet # another gap upcoming whose last tile is F then jump" script_part_2 = """ # if there's a gap just before jump target AND # we can continue jumping if we need, then jump NOT C J AND D J AND H J # however make sure NOT to walk a step when # being forced to jump would land us in gap NOT B T AND D T OR T J # if there's a gap right in front of us, jump NOT A T OR T J # fingers crossed RUN """ run_script(script_part_2)
0 [78, 79, 84, 32, 67, 32, 74, 10] 1 [65, 78, 68, 32, 68, 32, 74, 32, 10] 2 [65, 78, 68, 32, 72, 32, 74, 10] 3 [78, 79, 84, 32, 66, 32, 84, 32, 10] 4 [65, 78, 68, 32, 68, 32, 84, 32, 10] 5 [79, 82, 32, 84, 32, 74, 10] 6 [78, 79, 84, 32, 65, 32, 84, 32, 10] 7 [79, 82, 32, 84, 32, 74, 10] 8 [82, 85, 78, 10] retval = 1139528802