| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706 |
- "Usage: unparse.py <path to source file>"
- import sys
- import ast
- import tokenize
- import io
- import os
- # Large float and imaginary literals get turned into infinities in the AST.
- # We unparse those infinities to INFSTR.
- INFSTR = "1e" + repr(sys.float_info.max_10_exp + 1)
- def interleave(inter, f, seq):
- """Call f on each item in seq, calling inter() in between.
- """
- seq = iter(seq)
- try:
- f(next(seq))
- except StopIteration:
- pass
- else:
- for x in seq:
- inter()
- f(x)
- class Unparser:
- """Methods in this class recursively traverse an AST and
- output source code for the abstract syntax; original formatting
- is disregarded. """
- def __init__(self, tree, file = sys.stdout):
- """Unparser(tree, file=sys.stdout) -> None.
- Print the source for tree to file."""
- self.f = file
- self._indent = 0
- self.dispatch(tree)
- print("", file=self.f)
- self.f.flush()
- def fill(self, text = ""):
- "Indent a piece of text, according to the current indentation level"
- self.f.write("\n"+" "*self._indent + text)
- def write(self, text):
- "Append a piece of text to the current line."
- self.f.write(text)
- def enter(self):
- "Print ':', and increase the indentation."
- self.write(":")
- self._indent += 1
- def leave(self):
- "Decrease the indentation level."
- self._indent -= 1
- def dispatch(self, tree):
- "Dispatcher function, dispatching tree type T to method _T."
- if isinstance(tree, list):
- for t in tree:
- self.dispatch(t)
- return
- meth = getattr(self, "_"+tree.__class__.__name__)
- meth(tree)
- ############### Unparsing methods ######################
- # There should be one method per concrete grammar type #
- # Constructors should be grouped by sum type. Ideally, #
- # this would follow the order in the grammar, but #
- # currently doesn't. #
- ########################################################
- def _Module(self, tree):
- for stmt in tree.body:
- self.dispatch(stmt)
- # stmt
- def _Expr(self, tree):
- self.fill()
- self.dispatch(tree.value)
- def _Import(self, t):
- self.fill("import ")
- interleave(lambda: self.write(", "), self.dispatch, t.names)
- def _ImportFrom(self, t):
- self.fill("from ")
- self.write("." * t.level)
- if t.module:
- self.write(t.module)
- self.write(" import ")
- interleave(lambda: self.write(", "), self.dispatch, t.names)
- def _Assign(self, t):
- self.fill()
- for target in t.targets:
- self.dispatch(target)
- self.write(" = ")
- self.dispatch(t.value)
- def _AugAssign(self, t):
- self.fill()
- self.dispatch(t.target)
- self.write(" "+self.binop[t.op.__class__.__name__]+"= ")
- self.dispatch(t.value)
- def _AnnAssign(self, t):
- self.fill()
- if not t.simple and isinstance(t.target, ast.Name):
- self.write('(')
- self.dispatch(t.target)
- if not t.simple and isinstance(t.target, ast.Name):
- self.write(')')
- self.write(": ")
- self.dispatch(t.annotation)
- if t.value:
- self.write(" = ")
- self.dispatch(t.value)
- def _Return(self, t):
- self.fill("return")
- if t.value:
- self.write(" ")
- self.dispatch(t.value)
- def _Pass(self, t):
- self.fill("pass")
- def _Break(self, t):
- self.fill("break")
- def _Continue(self, t):
- self.fill("continue")
- def _Delete(self, t):
- self.fill("del ")
- interleave(lambda: self.write(", "), self.dispatch, t.targets)
- def _Assert(self, t):
- self.fill("assert ")
- self.dispatch(t.test)
- if t.msg:
- self.write(", ")
- self.dispatch(t.msg)
- def _Global(self, t):
- self.fill("global ")
- interleave(lambda: self.write(", "), self.write, t.names)
- def _Nonlocal(self, t):
- self.fill("nonlocal ")
- interleave(lambda: self.write(", "), self.write, t.names)
- def _Await(self, t):
- self.write("(")
- self.write("await")
- if t.value:
- self.write(" ")
- self.dispatch(t.value)
- self.write(")")
- def _Yield(self, t):
- self.write("(")
- self.write("yield")
- if t.value:
- self.write(" ")
- self.dispatch(t.value)
- self.write(")")
- def _YieldFrom(self, t):
- self.write("(")
- self.write("yield from")
- if t.value:
- self.write(" ")
- self.dispatch(t.value)
- self.write(")")
- def _Raise(self, t):
- self.fill("raise")
- if not t.exc:
- assert not t.cause
- return
- self.write(" ")
- self.dispatch(t.exc)
- if t.cause:
- self.write(" from ")
- self.dispatch(t.cause)
- def _Try(self, t):
- self.fill("try")
- self.enter()
- self.dispatch(t.body)
- self.leave()
- for ex in t.handlers:
- self.dispatch(ex)
- if t.orelse:
- self.fill("else")
- self.enter()
- self.dispatch(t.orelse)
- self.leave()
- if t.finalbody:
- self.fill("finally")
- self.enter()
- self.dispatch(t.finalbody)
- self.leave()
- def _ExceptHandler(self, t):
- self.fill("except")
- if t.type:
- self.write(" ")
- self.dispatch(t.type)
- if t.name:
- self.write(" as ")
- self.write(t.name)
- self.enter()
- self.dispatch(t.body)
- self.leave()
- def _ClassDef(self, t):
- self.write("\n")
- for deco in t.decorator_list:
- self.fill("@")
- self.dispatch(deco)
- self.fill("class "+t.name)
- self.write("(")
- comma = False
- for e in t.bases:
- if comma: self.write(", ")
- else: comma = True
- self.dispatch(e)
- for e in t.keywords:
- if comma: self.write(", ")
- else: comma = True
- self.dispatch(e)
- self.write(")")
- self.enter()
- self.dispatch(t.body)
- self.leave()
- def _FunctionDef(self, t):
- self.__FunctionDef_helper(t, "def")
- def _AsyncFunctionDef(self, t):
- self.__FunctionDef_helper(t, "async def")
- def __FunctionDef_helper(self, t, fill_suffix):
- self.write("\n")
- for deco in t.decorator_list:
- self.fill("@")
- self.dispatch(deco)
- def_str = fill_suffix+" "+t.name + "("
- self.fill(def_str)
- self.dispatch(t.args)
- self.write(")")
- if t.returns:
- self.write(" -> ")
- self.dispatch(t.returns)
- self.enter()
- self.dispatch(t.body)
- self.leave()
- def _For(self, t):
- self.__For_helper("for ", t)
- def _AsyncFor(self, t):
- self.__For_helper("async for ", t)
- def __For_helper(self, fill, t):
- self.fill(fill)
- self.dispatch(t.target)
- self.write(" in ")
- self.dispatch(t.iter)
- self.enter()
- self.dispatch(t.body)
- self.leave()
- if t.orelse:
- self.fill("else")
- self.enter()
- self.dispatch(t.orelse)
- self.leave()
- def _If(self, t):
- self.fill("if ")
- self.dispatch(t.test)
- self.enter()
- self.dispatch(t.body)
- self.leave()
- # collapse nested ifs into equivalent elifs.
- while (t.orelse and len(t.orelse) == 1 and
- isinstance(t.orelse[0], ast.If)):
- t = t.orelse[0]
- self.fill("elif ")
- self.dispatch(t.test)
- self.enter()
- self.dispatch(t.body)
- self.leave()
- # final else
- if t.orelse:
- self.fill("else")
- self.enter()
- self.dispatch(t.orelse)
- self.leave()
- def _While(self, t):
- self.fill("while ")
- self.dispatch(t.test)
- self.enter()
- self.dispatch(t.body)
- self.leave()
- if t.orelse:
- self.fill("else")
- self.enter()
- self.dispatch(t.orelse)
- self.leave()
- def _With(self, t):
- self.fill("with ")
- interleave(lambda: self.write(", "), self.dispatch, t.items)
- self.enter()
- self.dispatch(t.body)
- self.leave()
- def _AsyncWith(self, t):
- self.fill("async with ")
- interleave(lambda: self.write(", "), self.dispatch, t.items)
- self.enter()
- self.dispatch(t.body)
- self.leave()
- # expr
- def _Bytes(self, t):
- self.write(repr(t.s))
- def _Str(self, tree):
- self.write(repr(tree.s))
- def _JoinedStr(self, t):
- self.write("f")
- string = io.StringIO()
- self._fstring_JoinedStr(t, string.write)
- self.write(repr(string.getvalue()))
- def _FormattedValue(self, t):
- self.write("f")
- string = io.StringIO()
- self._fstring_FormattedValue(t, string.write)
- self.write(repr(string.getvalue()))
- def _fstring_JoinedStr(self, t, write):
- for value in t.values:
- meth = getattr(self, "_fstring_" + type(value).__name__)
- meth(value, write)
- def _fstring_Str(self, t, write):
- value = t.s.replace("{", "{{").replace("}", "}}")
- write(value)
- def _fstring_Constant(self, t, write):
- assert isinstance(t.value, str)
- value = t.value.replace("{", "{{").replace("}", "}}")
- write(value)
- def _fstring_FormattedValue(self, t, write):
- write("{")
- expr = io.StringIO()
- Unparser(t.value, expr)
- expr = expr.getvalue().rstrip("\n")
- if expr.startswith("{"):
- write(" ") # Separate pair of opening brackets as "{ {"
- write(expr)
- if t.conversion != -1:
- conversion = chr(t.conversion)
- assert conversion in "sra"
- write(f"!{conversion}")
- if t.format_spec:
- write(":")
- meth = getattr(self, "_fstring_" + type(t.format_spec).__name__)
- meth(t.format_spec, write)
- write("}")
- def _Name(self, t):
- self.write(t.id)
- def _write_constant(self, value):
- if isinstance(value, (float, complex)):
- self.write(repr(value).replace("inf", INFSTR))
- else:
- self.write(repr(value))
- def _Constant(self, t):
- value = t.value
- if isinstance(value, tuple):
- self.write("(")
- if len(value) == 1:
- self._write_constant(value[0])
- self.write(",")
- else:
- interleave(lambda: self.write(", "), self._write_constant, value)
- self.write(")")
- else:
- self._write_constant(t.value)
- def _NameConstant(self, t):
- self.write(repr(t.value))
- def _Num(self, t):
- # Substitute overflowing decimal literal for AST infinities.
- self.write(repr(t.n).replace("inf", INFSTR))
- def _List(self, t):
- self.write("[")
- interleave(lambda: self.write(", "), self.dispatch, t.elts)
- self.write("]")
- def _ListComp(self, t):
- self.write("[")
- self.dispatch(t.elt)
- for gen in t.generators:
- self.dispatch(gen)
- self.write("]")
- def _GeneratorExp(self, t):
- self.write("(")
- self.dispatch(t.elt)
- for gen in t.generators:
- self.dispatch(gen)
- self.write(")")
- def _SetComp(self, t):
- self.write("{")
- self.dispatch(t.elt)
- for gen in t.generators:
- self.dispatch(gen)
- self.write("}")
- def _DictComp(self, t):
- self.write("{")
- self.dispatch(t.key)
- self.write(": ")
- self.dispatch(t.value)
- for gen in t.generators:
- self.dispatch(gen)
- self.write("}")
- def _comprehension(self, t):
- if t.is_async:
- self.write(" async for ")
- else:
- self.write(" for ")
- self.dispatch(t.target)
- self.write(" in ")
- self.dispatch(t.iter)
- for if_clause in t.ifs:
- self.write(" if ")
- self.dispatch(if_clause)
- def _IfExp(self, t):
- self.write("(")
- self.dispatch(t.body)
- self.write(" if ")
- self.dispatch(t.test)
- self.write(" else ")
- self.dispatch(t.orelse)
- self.write(")")
- def _Set(self, t):
- assert(t.elts) # should be at least one element
- self.write("{")
- interleave(lambda: self.write(", "), self.dispatch, t.elts)
- self.write("}")
- def _Dict(self, t):
- self.write("{")
- def write_key_value_pair(k, v):
- self.dispatch(k)
- self.write(": ")
- self.dispatch(v)
- def write_item(item):
- k, v = item
- if k is None:
- # for dictionary unpacking operator in dicts {**{'y': 2}}
- # see PEP 448 for details
- self.write("**")
- self.dispatch(v)
- else:
- write_key_value_pair(k, v)
- interleave(lambda: self.write(", "), write_item, zip(t.keys, t.values))
- self.write("}")
- def _Tuple(self, t):
- self.write("(")
- if len(t.elts) == 1:
- elt = t.elts[0]
- self.dispatch(elt)
- self.write(",")
- else:
- interleave(lambda: self.write(", "), self.dispatch, t.elts)
- self.write(")")
- unop = {"Invert":"~", "Not": "not", "UAdd":"+", "USub":"-"}
- def _UnaryOp(self, t):
- self.write("(")
- self.write(self.unop[t.op.__class__.__name__])
- self.write(" ")
- self.dispatch(t.operand)
- self.write(")")
- binop = { "Add":"+", "Sub":"-", "Mult":"*", "MatMult":"@", "Div":"/", "Mod":"%",
- "LShift":"<<", "RShift":">>", "BitOr":"|", "BitXor":"^", "BitAnd":"&",
- "FloorDiv":"//", "Pow": "**"}
- def _BinOp(self, t):
- self.write("(")
- self.dispatch(t.left)
- self.write(" " + self.binop[t.op.__class__.__name__] + " ")
- self.dispatch(t.right)
- self.write(")")
- cmpops = {"Eq":"==", "NotEq":"!=", "Lt":"<", "LtE":"<=", "Gt":">", "GtE":">=",
- "Is":"is", "IsNot":"is not", "In":"in", "NotIn":"not in"}
- def _Compare(self, t):
- self.write("(")
- self.dispatch(t.left)
- for o, e in zip(t.ops, t.comparators):
- self.write(" " + self.cmpops[o.__class__.__name__] + " ")
- self.dispatch(e)
- self.write(")")
- boolops = {ast.And: 'and', ast.Or: 'or'}
- def _BoolOp(self, t):
- self.write("(")
- s = " %s " % self.boolops[t.op.__class__]
- interleave(lambda: self.write(s), self.dispatch, t.values)
- self.write(")")
- def _Attribute(self,t):
- self.dispatch(t.value)
- # Special case: 3.__abs__() is a syntax error, so if t.value
- # is an integer literal then we need to either parenthesize
- # it or add an extra space to get 3 .__abs__().
- if ((isinstance(t.value, ast.Num) and isinstance(t.value.n, int))
- or (isinstance(t.value, ast.Constant) and isinstance(t.value.value, int))):
- self.write(" ")
- self.write(".")
- self.write(t.attr)
- def _Call(self, t):
- self.dispatch(t.func)
- self.write("(")
- comma = False
- for e in t.args:
- if comma: self.write(", ")
- else: comma = True
- self.dispatch(e)
- for e in t.keywords:
- if comma: self.write(", ")
- else: comma = True
- self.dispatch(e)
- self.write(")")
- def _Subscript(self, t):
- self.dispatch(t.value)
- self.write("[")
- self.dispatch(t.slice)
- self.write("]")
- def _Starred(self, t):
- self.write("*")
- self.dispatch(t.value)
- # slice
- def _Ellipsis(self, t):
- self.write("...")
- def _Index(self, t):
- self.dispatch(t.value)
- def _Slice(self, t):
- if t.lower:
- self.dispatch(t.lower)
- self.write(":")
- if t.upper:
- self.dispatch(t.upper)
- if t.step:
- self.write(":")
- self.dispatch(t.step)
- def _ExtSlice(self, t):
- interleave(lambda: self.write(', '), self.dispatch, t.dims)
- # argument
- def _arg(self, t):
- self.write(t.arg)
- if t.annotation:
- self.write(": ")
- self.dispatch(t.annotation)
- # others
- def _arguments(self, t):
- first = True
- # normal arguments
- defaults = [None] * (len(t.args) - len(t.defaults)) + t.defaults
- for a, d in zip(t.args, defaults):
- if first:first = False
- else: self.write(", ")
- self.dispatch(a)
- if d:
- self.write("=")
- self.dispatch(d)
- # varargs, or bare '*' if no varargs but keyword-only arguments present
- if t.vararg or t.kwonlyargs:
- if first:first = False
- else: self.write(", ")
- self.write("*")
- if t.vararg:
- self.write(t.vararg.arg)
- if t.vararg.annotation:
- self.write(": ")
- self.dispatch(t.vararg.annotation)
- # keyword-only arguments
- if t.kwonlyargs:
- for a, d in zip(t.kwonlyargs, t.kw_defaults):
- if first:first = False
- else: self.write(", ")
- self.dispatch(a),
- if d:
- self.write("=")
- self.dispatch(d)
- # kwargs
- if t.kwarg:
- if first:first = False
- else: self.write(", ")
- self.write("**"+t.kwarg.arg)
- if t.kwarg.annotation:
- self.write(": ")
- self.dispatch(t.kwarg.annotation)
- def _keyword(self, t):
- if t.arg is None:
- self.write("**")
- else:
- self.write(t.arg)
- self.write("=")
- self.dispatch(t.value)
- def _Lambda(self, t):
- self.write("(")
- self.write("lambda ")
- self.dispatch(t.args)
- self.write(": ")
- self.dispatch(t.body)
- self.write(")")
- def _alias(self, t):
- self.write(t.name)
- if t.asname:
- self.write(" as "+t.asname)
- def _withitem(self, t):
- self.dispatch(t.context_expr)
- if t.optional_vars:
- self.write(" as ")
- self.dispatch(t.optional_vars)
- def roundtrip(filename, output=sys.stdout):
- with open(filename, "rb") as pyfile:
- encoding = tokenize.detect_encoding(pyfile.readline)[0]
- with open(filename, "r", encoding=encoding) as pyfile:
- source = pyfile.read()
- tree = compile(source, filename, "exec", ast.PyCF_ONLY_AST)
- Unparser(tree, output)
- def testdir(a):
- try:
- names = [n for n in os.listdir(a) if n.endswith('.py')]
- except OSError:
- print("Directory not readable: %s" % a, file=sys.stderr)
- else:
- for n in names:
- fullname = os.path.join(a, n)
- if os.path.isfile(fullname):
- output = io.StringIO()
- print('Testing %s' % fullname)
- try:
- roundtrip(fullname, output)
- except Exception as e:
- print(' Failed to compile, exception is %s' % repr(e))
- elif os.path.isdir(fullname):
- testdir(fullname)
- def main(args):
- if args[0] == '--testdir':
- for a in args[1:]:
- testdir(a)
- else:
- for a in args:
- roundtrip(a)
- if __name__=='__main__':
- main(sys.argv[1:])
|