lambdaV: Implement parser

At this point, I have no idea how well it actually works as I have
only performed basic sanity checks on it. But it seems to work.

lambdaV.py

@@ -0,0 +1,358 @@
+#!/usr/bin/env python3
+
+# Copyright 2022, Jan Danielzick (aka. BodgeMaster)
+#
+# This program is free software: you can redistribute it and/or modify it
+# under the terms of the GNU General Public License as published
+# by the Free Software Foundation, version 3.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied
+# warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
+# See the GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# version 3 along with this program.
+# If not, see https://www.gnu.org/licenses/gpl-3.0.en.html
+
+import re
+
+debug_mode = True
+
+cursor_up    = 'Λ'
+cursor_down  = 'V'
+cursor_left  = '<'
+cursor_right = '>'
+
+empty_field = [
+    " ________________ ",
+    "|                |",
+    "|                |",
+    "|                |",
+    "|                |",
+    "|                |",
+    "|                |",
+    "|                |",
+    "|                |",
+    " ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ "
+]
+
+wall  = '#'
+apple = 'ó'
+goal  = '$'
+
+cursor_position = [0, 0]
+
+
+def debug_message(text):
+    if debug_mode:
+        print("DEBUG: ", end="")
+        print(text)
+
+def condition_in_front_of_wall(inverted):
+    return False
+
+def condition_goal_reached(inverted):
+    return False
+
+def condition_on_apple(inverted):
+    return False
+
+def condition_facing_north(inverted):
+    return False
+
+def condition_facing_south(inverted):
+    return False
+
+def condition_facing_east(inverted):
+    return False
+
+def condition_facing_west(inverted):
+    return False
+
+# condition prefixed with !
+def modifier_not(condition):
+    return condition(true)
+
+# return value: "" or error message
+def command_step():
+    #TODO
+    return ""
+
+def command_left():
+    return ""
+
+def command_right():
+    return ""
+
+def command_repeat(number, parsed_code):
+    return ""
+
+def command_while(parsed_condition, parsed_code):
+    return ""
+
+def command_take():
+    return ""
+
+def command_if(parsed_condition, parsed_then_code, parsed_else_code):
+    return ""
+
+
+# returns -1 for generic error
+# returns -2 if end can’t be found
+def get_length_of_condition(code):
+    if len(code)==0:
+        return -1
+    if code[0] == '<':
+        length = 0
+        while length<len(code) and code[length]!='>':
+            length = length+1
+        if length==len(code) and code[-1]!='>':
+            return -2
+        return length+1
+    else:
+        return -1
+
+
+# returns -1 for generic error
+# returns -2 if end can’t be found
+def get_length_of_contained_code(code):
+    if len(code)==0:
+        return -1
+    if code[0] == '(':
+        length = 1
+        nesting_level = 1
+        while length<len(code) and nesting_level>0:
+            if code[length] == '(':
+                nesting_level = nesting_level+1
+            elif code[length] == ')':
+                nesting_level = nesting_level-1
+            length = length+1
+        if length==len(code) and nesting_level > 0:
+            return -2
+        return length
+    else:
+        return -1
+
+# clean up whitespace
+def format_code(code):
+    regex_pattern = re.compile(r'\s+')
+    code = " ".join(re.sub(regex_pattern, ' ', code).lower().split())
+    code = code.replace(" (", "(")
+    code = code.replace("( ", "(")
+    code = code.replace(" )", ")")
+    code = code.replace(") ", ")")
+    code = code.replace(" <", "<")
+    code = code.replace("< ", "<")
+    code = code.replace(" >", ">")
+    code = code.replace("< ", "<")
+    code = code.replace("! ", "!")
+    # add spaces back after closing parenthesis
+    code = code.replace(")", ") ")
+    code = code.replace(") )", "))")
+    code = code.replace(") )", "))")
+    if len(code)>0 and code[-1]==' ':
+        return code[:-1]
+    return code
+
+# returns [function, inverted?, error message]
+def parse_condition(condition_code, allowed_conditions):
+    inverted = False
+    condition = None
+    if len(condition_code)==0:
+        debug_message("  No condition specified")
+        return [condition, inverted, "Syntax error: No condition specified"]
+    if condition_code[0]=='!':
+        debug_message("  Condition is inverted")
+        inverted = True
+        condition_code = condition_code[1:]
+        if len(condition_code)==0:
+            debug_message("  No condition specified")
+            return [condition, inverted, "Syntax error: No condition specified"]
+
+    if condition_code == "in front of wall":
+        condition = condition_in_front_of_wall
+    elif condition_code == "goal reached":
+        condition = condition_goal_reached
+    elif condition_code == "on apple":
+        condition = condition_on_apple
+    elif condition_code == "facing north":
+        condition = condition_facing_north
+    elif condition_code == "facing south":
+        condition = condition_facing_south
+    elif condition_code == "facing east":
+        condition = condition_facing_east
+    elif condition_code == "facing west":
+        condition = condition_facing_west
+    else:
+        debug_message("  Unknown condition: "+condition_code)
+        return [condition, inverted, "Unknown condition: "+condition_code]
+    return [condition, inverted, ""]
+
+# returns [[functions], [arguments], -1, ""] or [[], [], error_position, "error message"]
+def parse_code(code, allowed_commands, allowed_conditions, unformatted_code=True):
+    if unformatted_code:
+        code = format_code(code)
+        debug_message("Formatted code: "+code)
+
+    parse_position = 0
+    parsed_code = [[], [], -1, ""]
+    while parse_position < len(code):
+        # find next command
+        next_space = code.find(' ', parse_position)
+        next_condition = code.find('<', parse_position)
+        if next_space == -1:
+            next_space = len(code)
+        if next_condition == -1:
+            next_condition = len(code)
+        # Are we working with a simple command?
+        if next_condition>next_space or next_condition==next_space:
+            next_command = code[parse_position:next_space]
+            debug_message("Next command is: "+next_command)
+
+            parsed_code[1].append(())
+            #TODO: add function to to the function list
+            if next_command == "step":
+                parsed_code[0].append(command_step)
+            elif next_command == "left":
+                parsed_code[0].append(command_left)
+            elif next_command == "right":
+                parsed_code[0].append(command_right)
+            elif next_command == "take":
+                parsed_code[0].append(command_take)
+            elif next_command == "repeat":
+                return [[], [], next_space, "Syntax error: Condition missing"]
+            elif next_command == "while":
+                return [[], [], next_space, "Syntax error: Condition missing"]
+            elif next_command == "if":
+                return [[], [], next_space, "Syntax error: Condition missing"]
+            elif next_command == "else":
+                return [[], [], parse_position, "Syntax error: Else without if statement"]
+            else:
+                #TODO: better error message (especially when special chars are detected inside next_command or a known command)
+                return [[], [], parse_position, "Unknown command: "+next_command]
+            parse_position = next_space+1
+        else:
+            debug_message("Found control structure...")
+            control_structure = code[parse_position:next_condition]
+            if control_structure == "repeat":
+                debug_message("  Type: repeat loop")
+                parse_position = next_condition
+                repetitions_length = get_length_of_condition(code[parse_position:])
+                if repetitions_length == -2:
+                    debug_message("  Mismatched <>")
+                    return [[], [], parse_position, "Syntax error: Cannot not find matching '>' for this '<'"]
+                repetitions = code[parse_position:parse_position+repetitions_length]
+                debug_message("  Number of repetitions: "+repetitions)
+                repetitions_int = 0
+                try:
+                    repetitions_int = int(repetitions[1:-1])
+                except ValueError:
+                    return [[], [], parse_position, "Not a valid number: "+repetitions[1:-1]]
+                parse_position = parse_position+len(repetitions)
+                contained_code_length = get_length_of_contained_code(code[parse_position:])
+                if contained_code_length == -1:
+                    debug_message("  Missing (")
+                    return [[], [], parse_position, "Syntax error: Cannot not find '(' for contained code"]
+                if contained_code_length == -2:
+                    debug_message("  Mismatched ()")
+                    return [[], [], parse_position, "Syntax error: Cannot not find matching ')' for this '('"]
+                contained_code = code[parse_position:parse_position+contained_code_length]
+                debug_message("  Contained code: "+contained_code)
+                parsed_contained_code = parse_code(contained_code[1:-1], allowed_commands, allowed_conditions, unformatted_code=False)
+                if not parsed_contained_code[2]==-1:
+                    debug_message("  Error while parsing contained code")
+                    return [[], [], parse_position+1+parsed_contained_code[2], parsed_contained_code[3]]
+                parse_position = parse_position+contained_code_length+1
+
+                parsed_code[1].append((repetitions_int, parsed_contained_code))
+                parsed_code[0].append(command_repeat)
+                continue
+            elif control_structure == "if":
+                debug_message("  Type: if statement")
+                parse_position = next_condition
+                condition_length = get_length_of_condition(code[parse_position:])
+                if condition_length == -2:
+                    debug_message("  Mismatched <>")
+                    return [[], [], parse_position, "Syntax error: Cannot not find matching '>' for this '<'"]
+                condition = code[parse_position:parse_position+condition_length]
+                debug_message("  Condition is: "+condition)
+                parsed_condition = parse_condition(condition[1:-1], allowed_conditions)
+                if not parsed_condition[2]=="":
+                    return [[], [], parse_position, parsed_condition[2]]
+                parse_position = parse_position+len(condition)
+                then_code_length = get_length_of_contained_code(code[parse_position:])
+                if then_code_length == -1:
+                    debug_message("  Missing (")
+                    return [[], [], parse_position, "Syntax error: Cannot not find '(' for then-code"]
+                if then_code_length == -2:
+                    debug_message("  Mismatched ()")
+                    return [[], [], parse_position, "Syntax error: Cannot not find matching ')' for this '('"]
+                then_code = code[parse_position:parse_position+then_code_length]
+                debug_message("  Then-code: "+then_code)
+                parsed_then_code = parse_code(then_code[1:-1], allowed_commands, allowed_conditions, unformatted_code=False)
+                if not parsed_then_code[2]==-1:
+                    debug_message("  Error while parsing then-code")
+                    return [[], [], parse_position+1+parsed_then_code[2], parsed_then_code[3]]
+                parse_position = parse_position+then_code_length+1
+                parsed_else_code = None
+                if parse_position<len(code)+4 and code[parse_position:parse_position+4]=="else":
+                    parse_position = parse_position+4
+                    else_code_length = get_length_of_contained_code(code[parse_position:])
+                    if else_code_length == -1:
+                        debug_message("  Missing (")
+                        return [[], [], parse_position, "Syntax error: Cannot not find '(' for else-code"]
+                    if else_code_length == -2:
+                        debug_message("  Mismatched ()")
+                        return [[], [], parse_position, "Syntax error: Cannot not find matching ')' for this '('"]
+                    else_code = code[parse_position:parse_position+else_code_length]
+                    debug_message("  Else-code: "+else_code)
+                    parsed_else_code = parse_code(else_code[1:-1], allowed_commands, allowed_conditions, unformatted_code=False)
+                    parse_position = parse_position+else_code_length+1
+
+                parsed_code[1].append((parsed_condition, parsed_then_code, parsed_else_code))
+                parsed_code[0].append(command_if)
+                continue
+            elif control_structure == "while":
+                debug_message("  Type: while loop")
+                parse_position = next_condition
+                condition_length = get_length_of_condition(code[parse_position:])
+                if condition_length == -2:
+                    debug_message("  Mismatched <>")
+                    return [[], [], parse_position, "Syntax error: Cannot not find matching '>' for this '<'"]
+                condition = code[parse_position:parse_position+condition_length]
+                debug_message("  Condition is: "+condition)
+                parsed_condition = parse_condition(condition[1:-1], allowed_conditions)
+                if not parsed_condition[2]=="":
+                    return [[], [], parse_position, parsed_condition[2]]
+                parse_position = parse_position+len(condition)
+                contained_code_length = get_length_of_contained_code(code[parse_position:])
+                if contained_code_length == -1:
+                    debug_message("  Missing (")
+                    return [[], [], parse_position, "Syntax error: Cannot not find '(' for contained code"]
+                if contained_code_length == -2:
+                    debug_message("  Mismatched ()")
+                    return [[], [], parse_position, "Syntax error: Cannot not find matching ')' for this '('"]
+                contained_code = code[parse_position:parse_position+contained_code_length]
+                debug_message("  Contained code: "+contained_code)
+                parsed_contained_code = parse_code(contained_code[1:-1], allowed_commands, allowed_conditions, unformatted_code=False)
+                if not parsed_contained_code[2]==-1:
+                    debug_message("  Error while parsing contained code")
+                    return [[], [], parse_position+1+parsed_contained_code[2], parsed_contained_code[3]]
+                parse_position = parse_position+contained_code_length+1
+
+                parsed_code[1].append((parsed_condition, parsed_contained_code))
+                parsed_code[0].append(command_while)
+            else:
+                debug_message("  Type: unknown control structure: "+control_structure)
+                #TODO: better error message (especially when special chars are detected inside control_structure or a known command)
+                return [[], [], parse_position, "Syntax error: Unknown control structure: "+control_structure]
+    return parsed_code
+
+# returns success or error message
+# Parsed code is the result of running the parser.
+def run_code(parsed_code):
+    return ""
+
+if __name__ == "__main__":
+    pass