def __init__(self, tokens):
self.lex_token = tokens
self.AST = SyntaxThreeNodeTree()
self.pos = 0
self.cur_node = self.AST
self.const_table = ConstTable()
self.name_table = NameTable()
self.var_table = VarNameTable()
self.bytecode = []
self.gobol_const_index = 0
self.gobol_name_index = 0
self.gobol_var_index = 0
self.code_line_number = 0
class Parser(object):
def __init__(self, tokens):
self.lex_token = tokens
self.AST = SyntaxThreeNodeTree()
self.pos = 0
self.cur_node = self.AST
self.const_table = ConstTable()
self.name_table = NameTable()
self.var_table = VarNameTable()
self.bytecode = []
self.gobol_const_index = 0
self.gobol_name_index = 0
self.gobol_var_index = 0
self.code_line_number = 0
# def next_token(self):
# pass
def main(self):
while self.lex_token[self.pos].value != "EOF":
if self.lex_token[self.pos].type == "VAR":
var_node = SyntaxTwoNodeTree()
self.var_node()
self.cur_node.set_value(var_node)
if self.lex_token[self.pos].type == "FUNC":
func_tmp_node = SyntaxTwoNodeTree()
func_block_node = SyntaxThreeNodeTree()
func_code_obj = FuncCodeObject()
pre_const_table = self.const_table
pre_name_table = self.name_table
pre_var_table = self.var_table
# self.const_table = func_code_obj.const_table
# self.name_table = func_code_obj.name_table
# self.var_table = func_code_obj.var_table
self.func_node(func_code_obj, func_tmp_node)
self.parser_block(func_block_node)
func_tmp_node.set_right(func_block_node)
self.cur_node.set_value(func_tmp_node)
self.const_table = pre_const_table
self.name_table = pre_name_table
self.var_table = pre_var_table
if self.lex_token[self.pos].type == "WHILE":
while_node = SyntaxTwoNodeTree()
while_block = SyntaxThreeNodeTree()
self.while_control_node(while_node)
self.parser_block(while_block)
while_node.set_right(while_block)
self.cur_node.set_value(while_node)
# self.cur_node = while_block
if self.lex_token[self.pos].type == "IF":
if_node = SyntaxTwoNodeTree()
if_block = SyntaxThreeNodeTree()
self.while_control_node(if_node)
self.parser_block(if_block)
if_node.set_right(if_block)
self.cur_node.set_value(if_node)
if self.lex_token[self.pos].type == "ASSIGN":
# self.pos += 1
assign_node = SyntaxTwoNodeTree()
self.assign_node(assign_node)
self.cur_node.set_value(assign_node)
if self.lex_token[self.pos].type == "RETURN":
self.return_value()
self.pos += 1
# print self.gobol_const_index
# print self.gobol_name_index
def parser_block(self, node):
while self.lex_token[self.pos].type != "RB_BRACKET":
if self.lex_token[self.pos].type == "VAR":
var_node = SyntaxTwoNodeTree()
self.var_node()
node.set_value(var_node)
if self.lex_token[self.pos].type == "WHILE":
while_node = SyntaxTwoNodeTree()
while_block = SyntaxThreeNodeTree()
self.while_control_node(while_node)
self.parser_block(while_block)
while_node.set_right(while_block)
node.set_value(while_node)
# self.cur_node = while_block
if self.lex_token[self.pos].type == "IF":
if_node = SyntaxTwoNodeTree()
if_block = SyntaxThreeNodeTree()
self.while_control_node(if_node)
self.parser_block(if_block)
if_node.set_right(if_block)
node.set_value(if_node)
if self.lex_token[self.pos].type == "ASSIGN":
# self.pos += 1
assign_node = SyntaxTwoNodeTree()
self.assign_node(assign_node)
node.set_value(assign_node)
if self.lex_token[self.pos].type == "RETURN":
self.return_value()
self.pos += 1
def var_node(self):
var_node = SyntaxTwoNodeTree()
while self.lex_token[self.pos].type != "ASSIGN":
self.pos += 1
var_node.setValue(self.lex_token[self.pos - 2].value, self.lex_token[self.pos - 1].value,
self.lex_token[self.pos + 1].value)
var_node.setExtraInfo("VAR:" + str(self.lex_token[self.pos - 2].value))
##SymbolTable
self.const_table.add(self.lex_token[self.pos + 1].value, self.gobol_const_index)
self.name_table.add(self.lex_token[self.pos - 1].value, self.gobol_name_index)
##bytecode
self.bytecode.append(str(self.code_line_number) + " LOAD_CONST " + str(self.gobol_const_index))
self.code_line_number += 1
self.gobol_const_index += 1
self.bytecode.append(str(self.code_line_number) + " STROKE_NAME " + str(self.gobol_name_index))
self.gobol_name_index += 1
self.code_line_number += 1
self.pos += 1
def func_node(self, codeObj, node):
var_num = 0
func_name = None
# func_tree_node = SyntaxTwoNodeTree()
while self.lex_token[self.pos].type != "LB_BRACKET":
if self.lex_token[self.pos].type == "LL_BRACKET":
func_name = self.lex_token[self.pos - 1].value
func_type = self.lex_token[self.pos - 2].value
while self.lex_token[self.pos].type != "RL_BRACKET":
if self.lex_token[self.pos].type == "MAOHAO":
##symbolTable
self.var_table.add(self.lex_token[self.pos - 1].value, self.gobol_var_index)
# print self.var_table.table[self.lex_token[self.pos - 1].value]
self.gobol_var_index += 1
var_num += 1
self.pos += 1
self.pos += 1
##add the node into AST
node.set_left(func_name)
##SymbolTable
self.const_table.add(codeObj, self.gobol_const_index)
self.name_table.add(func_name, self.gobol_name_index)
##bytecode
self.bytecode.append(str(self.code_line_number) + " LOAD_CONST " + str(self.gobol_const_index))
self.code_line_number += 1
self.gobol_const_index += 1
self.bytecode.append(str(self.code_line_number) + " STOKE_NAME " + str(self.gobol_name_index))
self.code_line_number += 1
self.gobol_name_index += 1
self.bytecode.append(str(self.code_line_number) + " MAKE_FUNCTION " + str(var_num))
self.code_line_number += 1
return codeObj
def assign_node(self, node):
##使用逆波兰表达式来计算
op_stack = []
res = []
var_tmp = self.lex_token[self.pos - 1].value
var_index = -1
node.set_left(var_tmp)
# 在符号表中寻找变量
for key in self.name_table.table:
if key == var_tmp:
var_index = self.name_table.table[key]
break
if isinstance(key, FuncCodeObject):
for i in key.name_table.table:
if i == var_tmp:
var_index = key.name_table.table[key]
break
# 运算符优先级
operator_priority = {'>': 0, '<': 0, '>=': 0, '<=': 0, '(': 0, ')': 0,
'+': 1, '-': 1, '*': 2, '/': 2, '++': 3, '--': 3, '!': 3}
##判断条件
while self.lex_token[self.pos].type != "FENGHAO":
# print op_stack
if self.check_var_and_digit(self.lex_token[self.pos].type):
res.append(self.lex_token[self.pos].value)
elif self.check_compute_operation(self.lex_token[self.pos].type):
if len(op_stack) != 0:
if operator_priority[self.lex_token[self.pos].value] > operator_priority[
op_stack[len(op_stack) - 1]]:
op_stack.append(self.lex_token[self.pos].value)
else:
tmp = op_stack.pop()
while operator_priority[tmp] > operator_priority[self.lex_token[self.pos].value]:
res.append(tmp)
tmp = op_stack.pop()
op_stack.append(tmp)
res.append(self.lex_token[self.pos].value)
else:
op_stack.append(self.lex_token[self.pos].value)
elif self.lex_token[self.pos].type == "LL_BRACKET":
op_stack.append(self.lex_token[self.pos].value)
elif self.lex_token[self.pos].type == "RL_BRACKET":
tmp = op_stack.pop()
while tmp != "(":
res.append(tmp)
tmp = op_stack.pop()
self.pos += 1
##最后把栈里的元素弹出
while op_stack != []:
res.append(op_stack.pop())
# print res
# 计算后缀表达式
result = self.postfixEval(res)
# print result
node.set_right(result)
##把变量的新的计算的值放进去 update SymbolTabel
for key in self.const_table.table:
if isinstance(key, FuncCodeObject):
for i in key.const_table.table:
if key.const_table.table[i] == var_index:
i = result
key.const_table.table[i] = var_index
break
else:
if self.const_table.table[key] == var_index:
key = result
self.const_table.table[key] = var_index
break
##bytecode
self.bytecode.append(str(self.code_line_number) + " LOAD_NAME " + str(var_index))
self.code_line_number += 1
self.bytecode.append(str(self.code_line_number) + " STOKE_NAME " + str(var_index))
self.code_line_number += 1
def while_control_node(self, node):
tmp_node = SyntaxTwoNodeTree()
while self.lex_token[self.pos].type != "LB_BRACKET":
self.pos += 1
if self.lex_token[self.pos].type == "LL_BRACKET":
while self.lex_token[self.pos].type != "RL_BRACKET":
self.pos += 1
if self.check_operation(self.lex_token[self.pos].type):
tmp_node.setExtraInfo("OP:" + str(self.lex_token[self.pos].type))
tmp_node.setValue(self.lex_token[self.pos].value, self.lex_token[self.pos - 1].value,
self.lex_token[self.pos + 1].value)
##SymbolTable
self.name_table.add(self.lex_token[self.pos - 1].value, self.gobol_name_index)
self.gobol_name_index += 1
self.bytecode.append(str(self.code_line_number) + " LOAD_NAME " + str(self.gobol_name_index))
self.code_line_number += 1
if self.lex_token[self.pos + 1].type == "VARIABLE":
self.name_table.add(self.lex_token[self.pos - 1].value, self.gobol_name_index)
self.gobol_name_index += 1
self.bytecode.append(str(self.code_line_number) + " LOAD_NAME " + str(self.gobol_name_index))
self.code_line_number += 1
elif self.lex_token[self.pos + 1].type == "DIGIT":
self.const_table.add(self.lex_token[self.pos + 1].value, self.gobol_const_index)
self.gobol_const_index += 1
self.bytecode.append(str(self.code_line_number) + " LOAD_CONST " + str(self.gobol_const_index))
self.code_line_number += 1
##bytecode
self.bytecode.append(str(self.code_line_number) + " COMPARE_OP " + str(op_index[self.lex_token[self.pos].value]))
self.code_line_number += 1
self.bytecode.append(str(self.code_line_number) + " JUMP_IF_FALSE " + "0")
self.code_line_number += 1
self.bytecode.append(str(self.code_line_number) + " POP_TOP " + "-1")
self.code_line_number += 1
self.pos += 1
node.set_left(tmp_node)
def if_control_node(self, node):
tmp_node = SyntaxTwoNodeTree()
print ">>>>>>>"
while self.lex_token[self.pos].type != "LB_BRACKET":
self.pos += 1
if self.lex_token[self.pos].type == "LL_BRACKET":
while self.lex_token[self.pos].type != "RL_BRACKET":
self.pos += 1
if self.check_operation(self.lex_token[self.pos].type):
tmp_node.setExtraInfo("OP:" + str(self.lex_token[self.pos].type))
tmp_node.setValue(self.lex_token[self.pos].value, self.lex_token[self.pos - 1].value,
self.lex_token[self.pos + 1].value)
##SymbolTable
self.bytecode.append(str(self.code_line_number) + " SETUP_LOOP " + "0")
self.code_line_number += 1
self.name_table.add(self.lex_token[self.pos - 1].value, self.gobol_name_index)
self.gobol_name_index += 1
self.bytecode.append(str(self.code_line_number) + " LOAD_NAME " + str(self.gobol_name_index))
self.code_line_number += 1
if self.lex_token[self.pos + 1].type == "VARIABLE":
self.name_table.add(self.lex_token[self.pos - 1].value, self.gobol_name_index)
self.gobol_name_index += 1
self.bytecode.append(str(self.code_line_number) + " LOAD_NAME " + str(self.gobol_name_index))
self.code_line_number += 1
elif self.lex_token[self.pos + 1].type == "DIGIT":
self.const_table.add(self.lex_token[self.pos + 1].value, self.gobol_const_index)
self.gobol_const_index += 1
self.bytecode.append(str(self.code_line_number) + " LOAD_CONST " + str(self.gobol_const_index))
self.code_line_number += 1
##bytecode
self.bytecode.append(str(self.code_line_number) + " COMPARE_OP " + str(op_index[self.lex_token[self.pos].value]))
self.code_line_number += 1
self.bytecode.append(str(self.code_line_number) + " JUMP_IF_FALSE " + "0")
self.code_line_number += 1
self.bytecode.append(str(self.code_line_number) + " POP_TOP " + "-1")
self.code_line_number += 1
self.pos += 1
node.set_left(tmp_node)
def return_value(self):
return_tmp = self.lex_token[self.pos+1].value
return_index = -1
for key in self.name_table.table:
if key == return_tmp:
return_index = self.name_table.table[key]
break
if isinstance(key, FuncCodeObject):
for i in key.name_table.table:
if i == return_tmp:
return_index = key.name_table.table[key]
break
##bytecode
self.bytecode.append(str(self.code_line_number)+" LOAD_NAME "+str(return_index))
self.code_line_number += 1
self.bytecode.append(str(self.code_line_number)+" RETURN_VALUE "+"-1")
self.code_line_number += 1
def check_operation(self, token):
return True if token == "LT" or token == "GT" or token == "GET" or token == "LET" else False
def check_compute_operation(self, token):
return True if token == "PLUS" or token == "MINUS" or token == "MUL" or token == "DIV" else False
def check_var_and_digit(self, token):
return True if token == "DIGIT" or token == "VARIABLE" else False
def postfixEval(self, tokens):
operandStack = Stack()
for i in tokens:
if i.isdigit():
operandStack.push(int(i))
else:
operand2 = operandStack.pop()
operand1 = operandStack.pop()
result = self.compute(operand1, operand2, i)
operandStack.push(result)
return operandStack.pop()
def compute(self, v1, v2, op):
if op == "+":
return v1 + v2
elif op == "-":
return v1 - v2
elif op == "*":
return v1 * v2
elif op == "/":
return v1 / v2