analyticalEngineFormat.py

import re
import collections

variableAssignmentRegex = re.compile(r"([a-zA-Z_]+)\s*=\s*(\d*)")
commandRegex = re.compile(r"^(:\w+:)?\s*([A-Z]+)\s*([-\w]+)?(')?")
jumpRegex = re.compile(r"(C[FB][\+\?])(:\w+:)")

variableInitFormat = "N{} {}"
lineerror = "Error in line [{}]: {}"

variables = collections.OrderedDict() # {varname: initialValue}
labels = collections.OrderedDict()    # {labelname: commandLine}
commands = []
romout = []

with open('input.txt') as rom:
  for lineno, line in enumerate(rom, 1):
    if line.startswith('#') or line.isspace():
      continue
    
    if '=' in line:
      varmatch = variableAssignmentRegex.match(line)

      if varmatch is None:
        print(lineerror.format(lineno, 'Varmatch failed'))
        continue

      variables[varmatch.group(1).strip()] = varmatch.group(2)
    else:
      if line.strip() in ['-', '+', '/', '*']:
        commands.append(line.strip())
        continue
      elif 'CB' in line or 'CF' in line:
        commands.append(line.strip())
        continue
      
      match = commandRegex.match(line)
      
      if match is None:
        print(lineerror.format(lineno, 'commatch failed'))
        continue

      if match.group(3) is not None:
        try:
          literal = int(match.group(3))
        except ValueError:
          literal = None
        
        if literal is not None: # replace literal with variable
          variables[str(literal)] = literal

      commands.append(line.strip())
      
for i, (varName, value) in enumerate(variables.items()):
  # add assignment to variables
  romout.append(variableInitFormat.format(i, value))

for index, command in enumerate(commands):
  if command in ['-', '+', '/', '*']:
    romout.append(command)
    continue

  if 'CB' in command or 'CF' in command:
    romout.append(command)
    continue

  match = commandRegex.match(command)

  if match is None:
    print("Error matching command - shouldn't happen idk\n", command, index)
    continue

  if match.group(3) in variables.keys(): # replace var with memloc
    romout.append("{} {}{}".format(match.group(2),
                                   list(variables.keys()).index(match.group(3)),
                                   match.group(4) if match.group(4) is not None else '')) # add the '
  else:
    romout.append(match.group(2))

  if match.group(1): # if there's a label
    labels[match.group(1)] = len(romout) - 1
# sort out labels
for lineno, line in enumerate(romout):
  if ':' not in line:
    continue
  
  match = jumpRegex.match(line)
  if match is None:
    print(lineerror.format(lineno, "jump regex error : " + line))
    continue
  
    # calculate how many instructions to go back
  if match.group(2) in labels.keys():
    if 'CB' in line:
      diff = lineno - labels[match.group(2)]
    else:
      diff = labels[match.group(2)] - lineno - 1
    romout[lineno] = ("{}{}".format(match.group(1), diff))

for line in romout:
  print(line)

input.txt

A = 0
B = 0
C = 0
D = 0

# if D = 0
+
L D
L -1

CF?:ifDZero:
CF+:DNotZero:

:ifDZero: B
CF+:END:

# A = B + C
:DNotZero: L B
L C
S A

/
# A = A / D
L A
L D
S A'

# if 100 - A < 0
-
L 100
L A
CF?:AGreaterThan100:
CF+:END:

:AGreaterThan100: L 100
L 0
S A

:END: B

output.ae

N0 0
N1 0
N2 0
N3 0
N4 -1
N5 100
N6 0
+
L 3
L 4
CF?1
CF+2
B
CF+15
L 1
L 2
S 0
/
L 0
L 3
S 0'
-
L 5
L 0
CF?1
CF+3
L 5
L 6
S 0
B