如何用Python中看似有效的s表达式修复解析错误？

将（字符串）s表达式解析为表示AST的嵌套数组/列表时，我遇到了一个意外错误。s表达式来自（SerAPI），但在我看来很好：

sexp = b'(Answer 3(ObjList((CoqGoal((fg_goals(((name 3)(ty(App(Ind(((Mutind(MPfile(DirPath((Id Logic)(Id Init)(Id Coq))))(DirPath())(Id eq))0)(Instance())))((Ind(((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id bool))0)(Instance())))(App(Const((Constant(MPfile(DirPath((Id Nat)(Id Init)(Id Coq))))(DirPath())(Id odd))(Instance())))((App(Construct((((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)2)(Instance())))((Construct((((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)1)(Instance())))))))(Construct((((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id bool))0)1)(Instance()))))))(hyp()))))(bg_goals())(shelved_goals())(given_up_goals()))))))\n'

当我尝试解析时，会收到以下错误消息：

train(policy,optimizer,env,gamma,nb_episodes=nb_episodes,ema_alpha=ema_alpha)
  File "main.py", line 88, in train
    state, reward, done, _ = env.step('Example test_oddb1: Nat.odd 1 = true.')
  File "/Users/korkejudith/home_simulation_research/coq-serapi-python/python_api/coq_env.py", line 120, in step
    state = self.state_embedder(state)
  File "/Users/korkejudith/home_simulation_research/coq-serapi-python/python_api/ai_mathematician.py", line 33, in __call__
    psexp = loads(str(sexp))
  File "/Users/korkejudith/miniconda3/envs/rltp/lib/python3.6/site-packages/sexpdata.py", line 243, in loads
    obj = parse(string, **kwds)
  File "/Users/korkejudith/miniconda3/envs/rltp/lib/python3.6/site-packages/sexpdata.py", line 675, in parse
    return Parser(string, **kwds).parse()
  File "/Users/korkejudith/miniconda3/envs/rltp/lib/python3.6/site-packages/sexpdata.py", line 655, in parse
    (i, sexp) = self.parse_sexp(0)
  File "/Users/korkejudith/miniconda3/envs/rltp/lib/python3.6/site-packages/sexpdata.py", line 641, in parse_sexp
    (i, subsexp) = self.parse_sexp(i + 1)
  File "/Users/korkejudith/miniconda3/envs/rltp/lib/python3.6/site-packages/sexpdata.py", line 642, in parse_sexp
    append(Quoted(subsexp[0]))
IndexError: list index out of range

我尝试了不同版本的字符串，但似乎都不起作用。有人知道会出什么问题吗

---

用不同版本复制错误的代码对我失败：

from sexpdata import loads, dumps

def buffer_test():
    print('buffer_test')
    sexp = ''' b'(Answer 3(ObjList((CoqGoal((fg_goals(((name 3)(ty(Prod(Name(Id n))(Ind(((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)(Instance())))(App(Ind(((Mutind(MPfile(DirPath((Id Logic)(Id Init)(Id Coq))))(DirPath())(Id eq))0)(Instance())))((Ind(((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)(Instance())))(App(Const((Constant(MPfile(DirPath((Id Nat)(Id Init)(Id Coq))))(DirPath())(Id add))(Instance())))((Construct((((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)1)(Instance())))(Rel 1)))(Rel 1)))))(hyp()))))(bg_goals())(shelved_goals())(given_up_goals()))))))\n'
    '''
    print(f'sexp = {sexp}')
    psexp = loads(sexp)

def buffer_str_test():
    print('buffer_str_test')
    sexp = str(''' b'(Answer 3(ObjList((CoqGoal((fg_goals(((name 3)(ty(Prod(Name(Id n))(Ind(((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)(Instance())))(App(Ind(((Mutind(MPfile(DirPath((Id Logic)(Id Init)(Id Coq))))(DirPath())(Id eq))0)(Instance())))((Ind(((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)(Instance())))(App(Const((Constant(MPfile(DirPath((Id Nat)(Id Init)(Id Coq))))(DirPath())(Id add))(Instance())))((Construct((((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)1)(Instance())))(Rel 1)))(Rel 1)))))(hyp()))))(bg_goals())(shelved_goals())(given_up_goals()))))))\n'
    ''')
    print(f'sexp = {sexp}')
    psexp = loads(sexp)

def str_str_test():
    print('str_str_test')
    sexp = str(''' (Answer 3(ObjList((CoqGoal((fg_goals(((name 3)(ty(Prod(Name(Id n))(Ind(((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)(Instance())))(App(Ind(((Mutind(MPfile(DirPath((Id Logic)(Id Init)(Id Coq))))(DirPath())(Id eq))0)(Instance())))((Ind(((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)(Instance())))(App(Const((Constant(MPfile(DirPath((Id Nat)(Id Init)(Id Coq))))(DirPath())(Id add))(Instance())))((Construct((((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)1)(Instance())))(Rel 1)))(Rel 1)))))(hyp()))))(bg_goals())(shelved_goals())(given_up_goals()))))))\n'
    ''')
    print(f'sexp = {sexp}')
    psexp = loads(sexp)

def str_test():
    print('str_test')
    sexp = ''' (Answer 3(ObjList((CoqGoal((fg_goals(((name 3)(ty(Prod(Name(Id n))(Ind(((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)(Instance())))(App(Ind(((Mutind(MPfile(DirPath((Id Logic)(Id Init)(Id Coq))))(DirPath())(Id eq))0)(Instance())))((Ind(((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)(Instance())))(App(Const((Constant(MPfile(DirPath((Id Nat)(Id Init)(Id Coq))))(DirPath())(Id add))(Instance())))((Construct((((Mutind(MPfile(DirPath((Id Datatypes)(Id Init)(Id Coq))))(DirPath())(Id nat))0)1)(Instance())))(Rel 1)))(Rel 1)))))(hyp()))))(bg_goals())(shelved_goals())(given_up_goals()))))))'
    '''
    print(f'sexp = {sexp}')
    psexp = loads(sexp)

if __name__ == '__main__':
    print('running main')
    buffer_test()
    buffer_str_test()
    str_str_test()
    str_test()
    print('sucessful main')

我还打开了这一期：

---

交叉张贴：

---

您需要将

sexp

修复为普通字符串。

---

暗示你的表达中有一个撇号。（调试将清楚地表明这一点。）它从何而来

  File "/Users/korkejudith/home_simulation_research/coq-serapi-python/python_api/ai_mathematician.py", line 33, in __call__
    psexp = loads(str(sexp))

---

这就是罪魁祸首

在不知道sexpdata的情况下，我觉得它可能与字符串类型有关。例如，

''b'

对我来说似乎是一个很大的困惑。嗯，好吧，所以我没有正确地将其转换为字符串。如何正确地将其转换为字符串？好的，所以

sexp.decode（'utf-8'）

似乎对我很有用。但是如何选择编码呢？我只是随机选择了一个。@CharlieParker您需要根据bytestring中可以出现的字符以及它们的编码方式（例如，国家字符）来选择它。如果除了ASCII字符之外什么都不能显示，那么您也可以使用

ASCII

.Hi，这里是sexpdata的维护者；问题确实存在于字符串引用中，撇号是引用列表和符号的lisp速记（

'（foo）

相当于

（quote（foo））

），因此看到

b'（foo）

sexpdata试图将其解析为符号

b

，然后是

'（foo）

，后跟

”

，并从中引发了一个神秘的错误。我有一个待合并的修复程序。实际上，写这篇评论让我觉得我们应该删除引用速记实现，因为它实际上属于LISP，而不是s表达式解析本身。

  File "/Users/korkejudith/home_simulation_research/coq-serapi-python/python_api/ai_mathematician.py", line 33, in __call__
    psexp = loads(str(sexp))