mirror of
https://github.com/davidhalter/jedi.git
synced 2025-12-07 06:24:27 +08:00
1706 lines
61 KiB
Python
1706 lines
61 KiB
Python
"""
|
|
Maintainer: David Halter <davidhalter88@gmail.com>
|
|
Version: 0.1
|
|
|
|
py_fuzzyparser parses python code, with the goal of a good representation of
|
|
the code within a tree structure. Variables, Classes and Functions are defined
|
|
within this tree structure, containing their exact locations in the code.
|
|
It is also a primary goal to work with code which contains syntax errors.
|
|
|
|
This behaviour may be used to refactor, modify, search and complete code.
|
|
|
|
To understand this code it is extremely important to understand the behaviour
|
|
of the python module 'tokenize'.
|
|
|
|
This original codebase of this parser, which has been refactored and heavily
|
|
changed, was programmed by Aaron Griffin <aaronmgriffin@gmail.com>.
|
|
|
|
**The structure of the following script:**
|
|
A Scope has
|
|
- imports (Import)
|
|
- subscopes (Scope, Class, Function, Flow)
|
|
- statements (Statement)
|
|
|
|
All these objects have `Name`s. `Call` and `Array` are used as detail objects
|
|
of a statement.
|
|
|
|
All those classes are being generated by PyFuzzyParser, which takes python text
|
|
as input and ignores just all the non-python stuff. Basically you could feed it
|
|
a perl script, and it should still work (which means throw no error.
|
|
"""
|
|
from _compatibility import (next, literal_eval, tokenize_func, BytesIO,
|
|
property, is_py3k, cleandoc, Python3Method)
|
|
|
|
import tokenize
|
|
import re
|
|
import keyword
|
|
import weakref
|
|
import os
|
|
|
|
import debug
|
|
|
|
|
|
class ParserError(Exception):
|
|
pass
|
|
|
|
|
|
def indent_block(text, indention=" "):
|
|
""" This function indents a text block with a default of four spaces """
|
|
temp = ''
|
|
while text and text[-1] == '\n':
|
|
temp += text[-1]
|
|
text = text[:-1]
|
|
lines = text.split('\n')
|
|
return '\n'.join(map(lambda s: indention + s, lines)) + temp
|
|
|
|
|
|
class PushBackIterator(object):
|
|
def __init__(self, iterator):
|
|
self.pushes = []
|
|
self.iterator = iterator
|
|
|
|
def push_back(self, value):
|
|
self.pushes.append(value)
|
|
|
|
def __iter__(self):
|
|
return self
|
|
|
|
def next(self):
|
|
""" Python 2 Compatibility """
|
|
return self.__next__()
|
|
|
|
def __next__(self):
|
|
if self.pushes:
|
|
return self.pushes.pop()
|
|
else:
|
|
return next(self.iterator)
|
|
|
|
|
|
class Base(object):
|
|
"""
|
|
This is just here to have an isinstance check, which is also used on
|
|
evaluate classes. But since they have sometimes a special type of
|
|
delegation, it is important for those classes to override this method.
|
|
|
|
I know that there is a chance to do such things with __instancecheck__, but
|
|
since Python 2.5 doesn't support it, I decided to do it this way.
|
|
"""
|
|
def isinstance(self, *cls):
|
|
return isinstance(self, cls)
|
|
|
|
|
|
class Simple(Base):
|
|
"""
|
|
The super class for Scope, Import, Name and Statement. Every object in
|
|
the parser tree inherits from this class.
|
|
"""
|
|
def __init__(self, start_pos, end_pos=(None, None)):
|
|
self.start_pos = start_pos
|
|
self.end_pos = end_pos
|
|
self.parent = lambda: None
|
|
|
|
@Python3Method
|
|
def get_parent_until(self, classes=(), reverse=False,
|
|
include_current=False):
|
|
""" Takes always the parent, until one class (not a Class) """
|
|
if type(classes) not in (tuple, list):
|
|
classes = (classes,)
|
|
scope = self
|
|
while not scope.parent() is None:
|
|
if classes and reverse != scope.isinstance(*classes):
|
|
if include_current:
|
|
return scope
|
|
break
|
|
scope = scope.parent()
|
|
return scope
|
|
|
|
def __repr__(self):
|
|
code = self.get_code().replace('\n', ' ')
|
|
return "<%s: %s@%s>" % \
|
|
(self.__class__.__name__, code, self.start_pos[0])
|
|
|
|
|
|
class Scope(Simple):
|
|
"""
|
|
Super class for the parser tree, which represents the state of a python
|
|
text file.
|
|
A Scope manages and owns its subscopes, which are classes and functions, as
|
|
well as variables and imports. It is used to access the structure of python
|
|
files.
|
|
|
|
:param start_pos: The position (line and column) of the scope.
|
|
:type start_pos: tuple(int, int)
|
|
:param docstr: The docstring for the current Scope.
|
|
:type docstr: str
|
|
"""
|
|
def __init__(self, start_pos, docstr=''):
|
|
super(Scope, self).__init__(start_pos)
|
|
self.subscopes = []
|
|
self.imports = []
|
|
self.statements = []
|
|
self.docstr = docstr
|
|
|
|
def add_scope(self, sub, decorators):
|
|
sub.parent = weakref.ref(self)
|
|
sub.decorators = decorators
|
|
for d in decorators:
|
|
# the parent is the same, because the decorator has not the scope
|
|
# of the function
|
|
d.parent = sub.parent
|
|
self.subscopes.append(sub)
|
|
return sub
|
|
|
|
def add_statement(self, stmt):
|
|
"""
|
|
Used to add a Statement or a Scope.
|
|
A statement would be a normal command (Statement) or a Scope (Flow).
|
|
"""
|
|
stmt.parent = weakref.ref(self)
|
|
self.statements.append(stmt)
|
|
return stmt
|
|
|
|
def add_docstr(self, string):
|
|
""" Clean up a docstring """
|
|
self.docstr = cleandoc(literal_eval(string))
|
|
|
|
def add_import(self, imp):
|
|
self.imports.append(imp)
|
|
imp.parent = weakref.ref(self)
|
|
|
|
def get_imports(self):
|
|
""" Gets also the imports within flow statements """
|
|
i = [] + self.imports
|
|
for s in self.statements:
|
|
if isinstance(s, Scope):
|
|
i += s.get_imports()
|
|
return i
|
|
|
|
def get_code(self, first_indent=False, indention=" "):
|
|
"""
|
|
:return: Returns the code of the current scope.
|
|
:rtype: str
|
|
"""
|
|
string = ""
|
|
if len(self.docstr) > 0:
|
|
string += '"""' + self.docstr + '"""\n'
|
|
for i in self.imports:
|
|
string += i.get_code()
|
|
for sub in self.subscopes:
|
|
string += sub.get_code(first_indent=True, indention=indention)
|
|
for stmt in self.statements:
|
|
string += stmt.get_code()
|
|
|
|
if first_indent:
|
|
string = indent_block(string, indention=indention)
|
|
return string
|
|
|
|
@Python3Method
|
|
def get_set_vars(self):
|
|
"""
|
|
Get all the names, that are active and accessible in the current
|
|
scope.
|
|
|
|
:return: list of Name
|
|
:rtype: list
|
|
"""
|
|
n = []
|
|
for stmt in self.statements:
|
|
try:
|
|
n += stmt.get_set_vars(True)
|
|
except TypeError:
|
|
n += stmt.get_set_vars()
|
|
|
|
# function and class names
|
|
n += [s.name for s in self.subscopes]
|
|
|
|
for i in self.imports:
|
|
if not i.star:
|
|
n += i.get_defined_names()
|
|
return n
|
|
|
|
def get_defined_names(self):
|
|
return [n for n in self.get_set_vars() \
|
|
if isinstance(n, Import) or len(n) == 1]
|
|
|
|
def is_empty(self):
|
|
"""
|
|
:return: True if there are no subscopes, imports and statements.
|
|
:rtype: bool
|
|
"""
|
|
return not (self.imports or self.subscopes or self.statements)
|
|
|
|
@Python3Method
|
|
def get_statement_for_position(self, pos):
|
|
for s in self.statements:
|
|
if isinstance(s, Flow):
|
|
p = s.get_statement_for_position(pos)
|
|
if s.next and not p:
|
|
p = s.next.get_statement_for_position(pos)
|
|
if p:
|
|
return p
|
|
elif s.start_pos <= pos < s.end_pos:
|
|
return s
|
|
|
|
for s in self.subscopes:
|
|
p = s.get_statement_for_position(pos)
|
|
if p:
|
|
return p
|
|
|
|
def __repr__(self):
|
|
try:
|
|
name = self.name
|
|
except AttributeError:
|
|
try:
|
|
name = self.command
|
|
except AttributeError:
|
|
name = self.path
|
|
|
|
return "<%s: %s@%s-%s>" % (self.__class__.__name__, name,
|
|
self.start_pos[0], self.end_pos[0])
|
|
|
|
|
|
class Module(Scope):
|
|
"""
|
|
The top scope, which is always a module.
|
|
"""
|
|
def __init__(self, path, docstr=''):
|
|
super(Module, self).__init__((1, 0), docstr)
|
|
self.path = path
|
|
self.global_vars = []
|
|
self._name = None
|
|
self.used_names = {}
|
|
self.temp_used_names = []
|
|
|
|
def add_global(self, name):
|
|
"""
|
|
Global means in these context a function (subscope) which has a global
|
|
statement.
|
|
This is only relevant for the top scope.
|
|
|
|
:param name: The name of the global.
|
|
:type name: Name
|
|
"""
|
|
self.global_vars.append(name)
|
|
# set no parent here, because globals are not defined in this scope.
|
|
|
|
def get_set_vars(self):
|
|
n = super(Module, self).get_set_vars()
|
|
n += self.global_vars
|
|
return n
|
|
|
|
def get_module_name(self):
|
|
""" This is used for the goto function. """
|
|
sep = (os.path.sep,) * 2
|
|
r = re.search(r'([^%s]+?)(%s__init__)?(\.py)?$' % sep, self.path)
|
|
string = r.group(1)
|
|
names = [(string, (0, 0))]
|
|
if not self._name:
|
|
self._name = Name(names, self.start_pos, self.end_pos, self)
|
|
return self._name
|
|
|
|
|
|
class Class(Scope):
|
|
"""
|
|
Used to store the parsed contents of a python class.
|
|
|
|
:param name: The Class name.
|
|
:type name: string
|
|
:param supers: The super classes of a Class.
|
|
:type supers: list
|
|
:param start_pos: The start position (line, column) of the class.
|
|
:type start_pos: tuple(int, int)
|
|
:param docstr: The docstring for the current Scope.
|
|
:type docstr: str
|
|
"""
|
|
def __init__(self, name, supers, start_pos, docstr=''):
|
|
super(Class, self).__init__(start_pos, docstr)
|
|
self.name = name
|
|
name.parent = weakref.ref(self)
|
|
self.supers = supers
|
|
for s in self.supers:
|
|
s.parent = weakref.ref(self)
|
|
self.decorators = []
|
|
|
|
def get_code(self, first_indent=False, indention=" "):
|
|
str = "\n".join('@' + stmt.get_code() for stmt in self.decorators)
|
|
str += 'class %s' % (self.name)
|
|
if len(self.supers) > 0:
|
|
sup = ','.join(stmt.code for stmt in self.supers)
|
|
str += '(%s)' % sup
|
|
str += ':\n'
|
|
str += super(Class, self).get_code(True, indention)
|
|
if self.is_empty():
|
|
str += "pass\n"
|
|
return str
|
|
|
|
|
|
class Function(Scope):
|
|
"""
|
|
Used to store the parsed contents of a python function.
|
|
|
|
:param name: The Function name.
|
|
:type name: string
|
|
:param params: The parameters (Statement) of a Function.
|
|
:type params: list
|
|
:param start_pos: The start position (line, column) the Function.
|
|
:type start_pos: tuple(int, int)
|
|
:param docstr: The docstring for the current Scope.
|
|
:type docstr: str
|
|
"""
|
|
def __init__(self, name, params, start_pos, annotation):
|
|
Scope.__init__(self, start_pos)
|
|
self.name = name
|
|
name.parent = weakref.ref(self)
|
|
self.params = params
|
|
for p in params:
|
|
p.parent = weakref.ref(self)
|
|
self.decorators = []
|
|
self.returns = []
|
|
self.is_generator = False
|
|
self.listeners = set() # not used here, but in evaluation.
|
|
|
|
if annotation is not None:
|
|
annotation.parent = weakref.ref(self)
|
|
self.annotation = annotation
|
|
|
|
def get_code(self, first_indent=False, indention=" "):
|
|
str = "\n".join('@' + stmt.get_code() for stmt in self.decorators)
|
|
params = ','.join([stmt.code for stmt in self.params])
|
|
str += "def %s(%s):\n" % (self.name, params)
|
|
str += super(Function, self).get_code(True, indention)
|
|
if self.is_empty():
|
|
str += "pass\n"
|
|
return str
|
|
|
|
def get_set_vars(self):
|
|
n = super(Function, self).get_set_vars()
|
|
for p in self.params:
|
|
try:
|
|
n.append(p.get_name())
|
|
except IndexError:
|
|
debug.warning("multiple names in param %s" % n)
|
|
return n
|
|
|
|
|
|
class Flow(Scope):
|
|
"""
|
|
Used to describe programming structure - flow statements,
|
|
which indent code, but are not classes or functions:
|
|
|
|
- for
|
|
- while
|
|
- if
|
|
- try
|
|
- with
|
|
|
|
Therefore statements like else, except and finally are also here,
|
|
they are now saved in the root flow elements, but in the next variable.
|
|
|
|
:param command: The flow command, if, while, else, etc.
|
|
:type command: str
|
|
:param inits: The initializations of a flow -> while 'statement'.
|
|
:type inits: list(Statement)
|
|
:param start_pos: Position (line, column) of the Flow statement.
|
|
:type start_pos: tuple(int, int)
|
|
:param set_vars: Local variables used in the for loop (only there).
|
|
:type set_vars: list
|
|
"""
|
|
def __init__(self, command, inits, start_pos, set_vars=None):
|
|
self.next = None
|
|
self.command = command
|
|
super(Flow, self).__init__(start_pos, '')
|
|
self._parent = None
|
|
# These have to be statements, because of with, which takes multiple.
|
|
self.inits = inits
|
|
for s in inits:
|
|
s.parent = weakref.ref(self)
|
|
if set_vars == None:
|
|
self.set_vars = []
|
|
else:
|
|
self.set_vars = set_vars
|
|
for s in self.set_vars:
|
|
s.parent = weakref.ref(self)
|
|
|
|
@property
|
|
def parent(self):
|
|
return self._parent
|
|
|
|
@parent.setter
|
|
def parent(self, value):
|
|
self._parent = value
|
|
if self.next:
|
|
self.next.parent = value
|
|
|
|
def get_code(self, first_indent=False, indention=" "):
|
|
if self.set_vars:
|
|
vars = ",".join(map(lambda x: x.get_code(), self.set_vars))
|
|
vars += ' in '
|
|
else:
|
|
vars = ''
|
|
|
|
stmts = []
|
|
for s in self.inits:
|
|
stmts.append(s.get_code(new_line=False))
|
|
stmt = ', '.join(stmts)
|
|
str = "%s %s%s:\n" % (self.command, vars, stmt)
|
|
str += super(Flow, self).get_code(True, indention)
|
|
if self.next:
|
|
str += self.next.get_code()
|
|
return str
|
|
|
|
def get_set_vars(self, is_internal_call=False):
|
|
"""
|
|
Get the names for the flow. This includes also a call to the super
|
|
class.
|
|
:param is_internal_call: defines an option for internal files to crawl\
|
|
through this class. Normally it will just call its superiors, to\
|
|
generate the output.
|
|
"""
|
|
if is_internal_call:
|
|
n = list(self.set_vars)
|
|
for s in self.inits:
|
|
n += s.set_vars
|
|
if self.next:
|
|
n += self.next.get_set_vars(is_internal_call)
|
|
n += super(Flow, self).get_set_vars()
|
|
return n
|
|
else:
|
|
return self.get_parent_until((Class, Function)).get_set_vars()
|
|
|
|
def get_imports(self):
|
|
i = super(Flow, self).get_imports()
|
|
if self.next:
|
|
i += self.next.get_imports()
|
|
return i
|
|
|
|
def set_next(self, next):
|
|
""" Set the next element in the flow, those are else, except, etc. """
|
|
if self.next:
|
|
return self.next.set_next(next)
|
|
else:
|
|
self.next = next
|
|
self.next.parent = self.parent
|
|
return next
|
|
|
|
|
|
class ForFlow(Flow):
|
|
"""
|
|
Used for the for loop, because there are two statement parts.
|
|
"""
|
|
def __init__(self, inits, start_pos, set_stmt, is_list_comp=False):
|
|
super(ForFlow, self).__init__('for', inits, start_pos,
|
|
set_stmt.used_vars)
|
|
self.set_stmt = set_stmt
|
|
self.is_list_comp = is_list_comp
|
|
|
|
|
|
class Import(Simple):
|
|
"""
|
|
Stores the imports of any Scopes.
|
|
|
|
>>> 1+1
|
|
2
|
|
|
|
:param start_pos: Position (line, column) of the Import.
|
|
:type start_pos: tuple(int, int)
|
|
:param namespace: The import, can be empty if a star is given
|
|
:type namespace: Name
|
|
:param alias: The alias of a namespace(valid in the current namespace).
|
|
:type alias: Name
|
|
:param from_ns: Like the namespace, can be equally used.
|
|
:type from_ns: Name
|
|
:param star: If a star is used -> from time import *.
|
|
:type star: bool
|
|
:param defunct: An Import is valid or not.
|
|
:type defunct: bool
|
|
"""
|
|
def __init__(self, start_pos, end_pos, namespace, alias=None, \
|
|
from_ns=None, star=False, relative_count=0, defunct=False):
|
|
super(Import, self).__init__(start_pos, end_pos)
|
|
|
|
self.namespace = namespace
|
|
self.alias = alias
|
|
self.from_ns = from_ns
|
|
for n in [namespace, alias, from_ns]:
|
|
if n:
|
|
n.parent = weakref.ref(self)
|
|
|
|
self.star = star
|
|
self.relative_count = relative_count
|
|
self.defunct = defunct
|
|
|
|
def get_code(self):
|
|
# in case one thing is None
|
|
alias = self.alias or ''
|
|
namespace = self.namespace or ''
|
|
from_ns = self.from_ns or ''
|
|
|
|
if self.alias:
|
|
ns_str = "%s as %s" % (namespace, alias)
|
|
else:
|
|
ns_str = str(namespace)
|
|
if self.from_ns or self.relative_count:
|
|
if self.star:
|
|
ns_str = '*'
|
|
dots = '.' * self.relative_count
|
|
return "from %s%s import %s\n" % (dots, from_ns, ns_str)
|
|
else:
|
|
return "import %s\n" % ns_str
|
|
|
|
def get_defined_names(self):
|
|
if self.defunct:
|
|
return []
|
|
if self.star:
|
|
return [self]
|
|
if self.alias:
|
|
return [self.alias]
|
|
if len(self.namespace) > 1:
|
|
o = self.namespace
|
|
n = Name([(o.names[0], o.start_pos)], o.start_pos, o.end_pos,
|
|
parent=o.parent())
|
|
return [n]
|
|
else:
|
|
return [self.namespace]
|
|
|
|
def get_all_import_names(self):
|
|
n = []
|
|
if self.from_ns:
|
|
n.append(self.from_ns)
|
|
if self.namespace:
|
|
n.append(self.namespace)
|
|
if self.alias:
|
|
n.append(self.alias)
|
|
return n
|
|
|
|
|
|
class Statement(Simple):
|
|
"""
|
|
This is the class for all the possible statements. Which means, this class
|
|
stores pretty much all the Python code, except functions, classes, imports,
|
|
and flow functions like if, for, etc.
|
|
|
|
:param code: The full code of a statement. This is import, if one wants \
|
|
to execute the code at some level.
|
|
:param code: str
|
|
:param set_vars: The variables which are defined by the statement.
|
|
:param set_vars: str
|
|
:param used_funcs: The functions which are used by the statement.
|
|
:param used_funcs: str
|
|
:param used_vars: The variables which are used by the statement.
|
|
:param used_vars: str
|
|
:param token_list: Token list which is also peppered with Name.
|
|
:param token_list: list
|
|
:param start_pos: Position (line, column) of the Statement.
|
|
:type start_pos: tuple(int, int)
|
|
"""
|
|
def __init__(self, code, set_vars, used_funcs, used_vars, token_list,
|
|
start_pos, end_pos):
|
|
super(Statement, self).__init__(start_pos, end_pos)
|
|
self.code = code
|
|
self.used_funcs = used_funcs
|
|
self.used_vars = used_vars
|
|
self.token_list = token_list
|
|
for s in set_vars + used_funcs + used_vars:
|
|
s.parent = weakref.ref(self)
|
|
self.set_vars = self._remove_executions_from_set_vars(set_vars)
|
|
|
|
# cache
|
|
self._assignment_calls = None
|
|
self._assignment_details = None
|
|
# this is important for other scripts
|
|
self._assignment_calls_calculated = False
|
|
|
|
def _remove_executions_from_set_vars(self, set_vars):
|
|
"""
|
|
Important mainly for assosiative arrays:
|
|
|
|
>>> a = 3
|
|
>>> b = {}
|
|
>>> b[a] = 3
|
|
|
|
`a` is in this case not a set_var, it is used to index the dict.
|
|
"""
|
|
|
|
if not set_vars:
|
|
return set_vars
|
|
result = set(set_vars)
|
|
last = None
|
|
in_execution = 0
|
|
for tok in self.token_list:
|
|
if isinstance(tok, Name):
|
|
if tok not in result:
|
|
break
|
|
if in_execution:
|
|
result.remove(tok)
|
|
elif isinstance(tok, tuple):
|
|
tok = tok[1]
|
|
if tok in ['(', '['] and isinstance(last, Name):
|
|
in_execution += 1
|
|
elif tok in [')', ']'] and in_execution > 0:
|
|
in_execution -= 1
|
|
last = tok
|
|
return list(result)
|
|
|
|
def get_code(self, new_line=True):
|
|
if new_line:
|
|
return self.code + '\n'
|
|
else:
|
|
return self.code
|
|
|
|
def get_set_vars(self):
|
|
""" Get the names for the statement. """
|
|
return list(self.set_vars)
|
|
|
|
@property
|
|
def assignment_details(self):
|
|
if self._assignment_details is None:
|
|
# normally, this calls sets this variable
|
|
self.get_assignment_calls()
|
|
# it may not have been set by get_assignment_calls -> just use an empty
|
|
# array
|
|
return self._assignment_details or []
|
|
|
|
def is_global(self):
|
|
# first keyword of the first token is global -> must be a global
|
|
return str(self.token_list[0]) == "global"
|
|
|
|
def get_assignment_calls(self):
|
|
"""
|
|
This is not done in the main parser, because it might be slow and
|
|
most of the statements won't need this data anyway. This is something
|
|
'like' a lazy execution.
|
|
|
|
This is not really nice written, sorry for that. If you plan to replace
|
|
it and make it nicer, that would be cool :-)
|
|
"""
|
|
if self._assignment_calls_calculated:
|
|
return self._assignment_calls
|
|
self._assignment_details = []
|
|
result = Array(self.start_pos, Array.NOARRAY, self)
|
|
top = result
|
|
level = 0
|
|
is_chain = False
|
|
close_brackets = False
|
|
|
|
tok_iter = enumerate(self.token_list)
|
|
for i, tok_temp in tok_iter:
|
|
#print 'tok', tok_temp, result
|
|
if isinstance(tok_temp, ListComprehension):
|
|
result.add_to_current_field(tok_temp)
|
|
continue
|
|
try:
|
|
token_type, tok, start_pos = tok_temp
|
|
except TypeError:
|
|
# the token is a Name, which has already been parsed
|
|
tok = tok_temp
|
|
token_type = None
|
|
start_pos = tok.start_pos
|
|
except ValueError:
|
|
debug.warning("unkown value, shouldn't happen",
|
|
tok_temp, type(tok_temp))
|
|
raise
|
|
else:
|
|
if tok in ['return', 'yield'] or level == 0 and \
|
|
'=' in tok and not tok in ['>=', '<=', '==', '!=']:
|
|
# This means, there is an assignment here.
|
|
|
|
# Add assignments, which can be more than one
|
|
self._assignment_details.append((tok, top))
|
|
# All these calls wouldn't be important if nonlocal would
|
|
# exist. -> Initialize the first item again.
|
|
result = Array(start_pos, Array.NOARRAY, self)
|
|
top = result
|
|
level = 0
|
|
close_brackets = False
|
|
is_chain = False
|
|
continue
|
|
elif tok == 'as':
|
|
next(tok_iter)
|
|
continue
|
|
|
|
brackets = {'(': Array.TUPLE, '[': Array.LIST, '{': Array.SET}
|
|
is_call = lambda: result.__class__ == Call
|
|
is_call_or_close = lambda: is_call() or close_brackets
|
|
|
|
is_literal = token_type in [tokenize.STRING, tokenize.NUMBER]
|
|
if isinstance(tok, Name) or is_literal:
|
|
c_type = Call.NAME
|
|
if is_literal:
|
|
tok = literal_eval(tok)
|
|
if token_type == tokenize.STRING:
|
|
c_type = Call.STRING
|
|
elif token_type == tokenize.NUMBER:
|
|
c_type = Call.NUMBER
|
|
|
|
if is_chain:
|
|
call = Call(tok, c_type, start_pos, parent=result)
|
|
result = result.set_next_chain_call(call)
|
|
is_chain = False
|
|
close_brackets = False
|
|
else:
|
|
if close_brackets:
|
|
result = result.parent()
|
|
close_brackets = False
|
|
if result.__class__ == Call:
|
|
result = result.parent()
|
|
close_brackets = False
|
|
call = Call(tok, c_type, start_pos, parent=result)
|
|
result.add_to_current_field(call)
|
|
result = call
|
|
elif tok in brackets.keys(): # brackets
|
|
level += 1
|
|
if is_call_or_close():
|
|
result = Array(start_pos, brackets[tok], parent=result)
|
|
result = result.parent().add_execution(result)
|
|
close_brackets = False
|
|
else:
|
|
result = Array(start_pos, brackets[tok], parent=result)
|
|
result.parent().add_to_current_field(result)
|
|
elif tok == ':':
|
|
if is_call_or_close():
|
|
result = result.parent()
|
|
close_brackets = False
|
|
if result.type == Array.LIST: # [:] lookups
|
|
result.add_to_current_field(tok)
|
|
else:
|
|
result.add_dictionary_key()
|
|
elif tok == '.':
|
|
if close_brackets and result.parent() != top:
|
|
# only get out of the array, if it is a array execution
|
|
result = result.parent()
|
|
close_brackets = False
|
|
is_chain = True
|
|
elif tok == ',':
|
|
while is_call_or_close():
|
|
result = result.parent()
|
|
close_brackets = False
|
|
result.add_field()
|
|
# important - it cannot be empty anymore
|
|
if result.type == Array.NOARRAY:
|
|
result.type = Array.TUPLE
|
|
elif tok in [')', '}', ']']:
|
|
while is_call_or_close():
|
|
result = result.parent()
|
|
close_brackets = False
|
|
if tok == '}' and not len(result):
|
|
# this is a really special case - empty brackets {} are
|
|
# always dictionaries and not sets.
|
|
result.type = Array.DICT
|
|
level -= 1
|
|
close_brackets = True
|
|
else:
|
|
while is_call_or_close():
|
|
result = result.parent()
|
|
close_brackets = False
|
|
result.add_to_current_field(tok)
|
|
|
|
if level != 0:
|
|
debug.warning("Brackets don't match: %s."
|
|
"This is not normal behaviour." % level)
|
|
|
|
self._assignment_calls_calculated = True
|
|
self._assignment_calls = top
|
|
return top
|
|
|
|
|
|
class Param(Statement):
|
|
"""
|
|
The class which shows definitions of params of classes and functions.
|
|
But this is not to define function calls.
|
|
"""
|
|
def __init__(self, code, set_vars, used_funcs, used_vars, token_list,
|
|
start_pos, end_pos):
|
|
super(Param, self).__init__(code, set_vars, used_funcs,
|
|
used_vars, token_list, start_pos, end_pos)
|
|
|
|
# this is defined by the parser later on, not at the initialization
|
|
# it is the position in the call (first argument, second...)
|
|
self.position_nr = None
|
|
self.is_generated = False
|
|
self.annotation_stmt = None
|
|
|
|
def add_annotation(self, annotation_stmt):
|
|
annotation_stmt.parent = weakref.ref(self)
|
|
self.annotation_stmt = annotation_stmt
|
|
|
|
def get_name(self):
|
|
""" get the name of the param """
|
|
n = self.set_vars or self.used_vars
|
|
if len(n) > 1:
|
|
raise IndexError("Multiple param names (%s)." % n)
|
|
return n[0]
|
|
|
|
|
|
class Call(Base):
|
|
"""
|
|
`Call` contains a call, e.g. `foo.bar` and owns the executions of those
|
|
calls, which are `Array`s.
|
|
"""
|
|
NAME = 1
|
|
NUMBER = 2
|
|
STRING = 3
|
|
|
|
def __init__(self, name, type, start_pos, parent_stmt=None, parent=None):
|
|
self.name = name
|
|
# parent is not the oposite of next. The parent of c: a = [b.c] would
|
|
# be an array.
|
|
self.parent = weakref.ref(parent) if parent is not None else None
|
|
self.type = type
|
|
self.start_pos = start_pos
|
|
|
|
self.next = None
|
|
self.execution = None
|
|
self._parent_stmt = weakref.ref(parent_stmt) if parent_stmt else None
|
|
|
|
@property
|
|
def parent_stmt(self):
|
|
if self._parent_stmt is not None:
|
|
return self._parent_stmt
|
|
elif self.parent:
|
|
return self.parent().parent_stmt
|
|
else:
|
|
return lambda: None
|
|
|
|
@parent_stmt.setter
|
|
def parent_stmt(self, value):
|
|
self._parent_stmt = value
|
|
|
|
def set_next_chain_call(self, call):
|
|
""" Adds another part of the statement"""
|
|
self.next = call
|
|
call.parent = self.parent
|
|
return call
|
|
|
|
def add_execution(self, call):
|
|
"""
|
|
An execution is nothing else than brackets, with params in them, which
|
|
shows access on the internals of this name.
|
|
"""
|
|
self.execution = call
|
|
# there might be multiple executions, like a()[0], in that case, they
|
|
# have the same parent. Otherwise it's not possible to parse proper.
|
|
if self.parent().execution == self:
|
|
call.parent = self.parent
|
|
else:
|
|
call.parent = weakref.ref(self)
|
|
return call
|
|
|
|
def generate_call_path(self):
|
|
""" Helps to get the order in which statements are executed. """
|
|
# TODO include previous nodes? As an option?
|
|
try:
|
|
for name_part in self.name.names:
|
|
yield name_part
|
|
except AttributeError:
|
|
yield self
|
|
if self.execution is not None:
|
|
for y in self.execution.generate_call_path():
|
|
yield y
|
|
if self.next is not None:
|
|
for y in self.next.generate_call_path():
|
|
yield y
|
|
|
|
def __repr__(self):
|
|
return "<%s: %s>" % \
|
|
(self.__class__.__name__, self.name)
|
|
|
|
|
|
class Array(Call):
|
|
"""
|
|
Describes the different python types for an array, but also empty
|
|
statements. In the Python syntax definitions this type is named 'atom'.
|
|
http://docs.python.org/py3k/reference/grammar.html
|
|
Array saves sub-arrays as well as normal operators and calls to methods.
|
|
|
|
:param array_type: The type of an array, which can be one of the constants\
|
|
below.
|
|
:type array_type: int
|
|
"""
|
|
NOARRAY = None
|
|
TUPLE = 'tuple'
|
|
LIST = 'list'
|
|
DICT = 'dict'
|
|
SET = 'set'
|
|
|
|
def __init__(self, start_pos, arr_type=NOARRAY, parent_stmt=None,
|
|
parent=None, values=None):
|
|
super(Array, self).__init__(None, arr_type, start_pos, parent_stmt,
|
|
parent)
|
|
|
|
self.values = values if values else []
|
|
self.keys = []
|
|
|
|
def add_field(self):
|
|
"""
|
|
Just add a new field to the values.
|
|
|
|
Each value has a sub-array, because there may be different tokens in
|
|
one array.
|
|
"""
|
|
self.values.append([])
|
|
|
|
def add_to_current_field(self, tok):
|
|
""" Adds a token to the latest field (in content). """
|
|
if not self.values:
|
|
# An empty round brace is just a tuple, filled it is unnecessary.
|
|
if self.type == Array.TUPLE:
|
|
self.type = Array.NOARRAY
|
|
# Add the first field, this is done here, because if nothing
|
|
# gets added, the list is empty, which is also needed sometimes.
|
|
self.values.append([])
|
|
self.values[-1].append(tok)
|
|
|
|
def add_dictionary_key(self):
|
|
"""
|
|
Only used for dictionaries, automatically adds the tokens added by now
|
|
from the values to keys, because the parser works this way.
|
|
"""
|
|
self.type = Array.DICT
|
|
self.keys.append(self.values.pop())
|
|
self.values.append([])
|
|
|
|
def get_only_subelement(self):
|
|
"""
|
|
Returns the only element that an array contains. If it contains
|
|
more than one element, raise an exception.
|
|
"""
|
|
if len(self.values) != 1 or len(self.values[0]) != 1:
|
|
raise AttributeError("More than one value found")
|
|
return self.values[0][0]
|
|
|
|
@staticmethod
|
|
def is_type(instance, *types):
|
|
"""
|
|
This is not only used for calls on the actual object, but for
|
|
ducktyping, to invoke this function with anything as `self`.
|
|
"""
|
|
if isinstance(instance, Array):
|
|
if instance.type in types:
|
|
return True
|
|
return False
|
|
|
|
def __len__(self):
|
|
return len(self.values)
|
|
|
|
def __getitem__(self, key):
|
|
return self.values[key]
|
|
|
|
def __iter__(self):
|
|
if self.type == self.DICT:
|
|
return iter(zip(self.keys, self.values))
|
|
else:
|
|
return iter(self.values)
|
|
|
|
def __repr__(self):
|
|
if self.type == self.NOARRAY:
|
|
type = 'noarray'
|
|
else:
|
|
type = self.type
|
|
return "<%s: %s%s>" % (self.__class__.__name__, type, self.values)
|
|
|
|
|
|
class NamePart(str):
|
|
"""
|
|
A string. Sometimes it is important to know if the string belongs to a name
|
|
or not.
|
|
"""
|
|
def __new__(cls, s, start_pos):
|
|
self = super(NamePart, cls).__new__(cls, s)
|
|
self.start_pos = start_pos
|
|
return self
|
|
|
|
@property
|
|
def end_pos(self):
|
|
return self.start_pos[0], self.start_pos[1] + len(self)
|
|
|
|
|
|
class Name(Simple):
|
|
"""
|
|
Used to define names in python.
|
|
Which means the whole namespace/class/function stuff.
|
|
So a name like "module.class.function"
|
|
would result in an array of [module, class, function]
|
|
"""
|
|
def __init__(self, names, start_pos, end_pos, parent=None):
|
|
super(Name, self).__init__(start_pos, end_pos)
|
|
self.names = tuple(n if isinstance(n, NamePart) else NamePart(*n)
|
|
for n in names)
|
|
if parent is not None:
|
|
self.parent = weakref.ref(parent)
|
|
|
|
def get_code(self):
|
|
""" Returns the names in a full string format """
|
|
return ".".join(self.names)
|
|
|
|
def __str__(self):
|
|
return self.get_code()
|
|
|
|
def __len__(self):
|
|
return len(self.names)
|
|
|
|
|
|
class ListComprehension(object):
|
|
""" Helper class for list comprehensions """
|
|
def __init__(self, stmt, middle, input):
|
|
self.stmt = stmt
|
|
self.middle = middle
|
|
self.input = input
|
|
|
|
def __repr__(self):
|
|
return "<%s: %s>" % \
|
|
(self.__class__.__name__, self.get_code())
|
|
|
|
def get_code(self):
|
|
statements = self.stmt, self.middle, self.input
|
|
code = [s.get_code().replace('\n', '') for s in statements]
|
|
return "%s for %s in %s" % tuple(code)
|
|
|
|
|
|
class PyFuzzyParser(object):
|
|
"""
|
|
This class is used to parse a Python file, it then divides them into a
|
|
class structure of different scopes.
|
|
|
|
:param code: The codebase for the parser.
|
|
:type code: str
|
|
:param user_position: The line/column, the user is currently on.
|
|
:type user_position: tuple(int, int)
|
|
"""
|
|
def __init__(self, code, module_path=None, user_position=None,
|
|
no_docstr=False):
|
|
self.user_position = user_position
|
|
self.user_scope = None
|
|
self.user_stmt = None
|
|
self.code = code + '\n' # end with \n, because the parser needs it
|
|
self.no_docstr = no_docstr
|
|
if is_py3k:
|
|
self.code = self.code.encode()
|
|
|
|
# initialize global Scope
|
|
self.module = Module(module_path)
|
|
self.scope = self.module
|
|
self.current = (None, None)
|
|
|
|
# Stuff to fix tokenize errors. The parser is pretty good in tolerating
|
|
# any errors of tokenize and just parse ahead.
|
|
self._tokenize_start_pos = 0
|
|
self._line_of_tokenize_restart = 0
|
|
|
|
self.parse()
|
|
|
|
# delete code again, only the parser needs it
|
|
del self.code
|
|
|
|
def __repr__(self):
|
|
return "<%s: %s>" % (self.__class__.__name__, self.module)
|
|
|
|
@property
|
|
def start_pos(self):
|
|
return (self._line_of_tokenize_restart + self._tokenize_start_pos[0],
|
|
self._tokenize_start_pos[1])
|
|
|
|
@property
|
|
def end_pos(self):
|
|
return (self._line_of_tokenize_restart + self._tokenize_end_pos[0],
|
|
self._tokenize_end_pos[1])
|
|
|
|
def _check_user_stmt(self, simple):
|
|
if not self.user_position:
|
|
return
|
|
# the position is right
|
|
if simple.start_pos <= self.user_position <= simple.end_pos:
|
|
if self.user_stmt is not None:
|
|
# if there is already a user position (another import, because
|
|
# imports are splitted) the names are checked.
|
|
for n in simple.get_defined_names():
|
|
if n.start_pos < self.user_position <= n.end_pos:
|
|
self.user_stmt = simple
|
|
else:
|
|
self.user_stmt = simple
|
|
|
|
def _parsedotname(self, pre_used_token=None):
|
|
"""
|
|
The dot name parser parses a name, variable or function and returns
|
|
their names.
|
|
|
|
:return: Tuple of Name, token_type, nexttoken.
|
|
:rtype: tuple(Name, int, str)
|
|
"""
|
|
def append(el):
|
|
names.append(el)
|
|
self.module.temp_used_names.append(el[0])
|
|
|
|
names = []
|
|
if pre_used_token is None:
|
|
token_type, tok = self.next()
|
|
if token_type != tokenize.NAME and tok != '*':
|
|
return ([], token_type, tok, self.start_pos)
|
|
else:
|
|
token_type, tok = pre_used_token
|
|
|
|
if token_type != tokenize.NAME and tok != '*':
|
|
# token maybe a name or star
|
|
return (None, token_type, tok)
|
|
|
|
append((tok, self.start_pos))
|
|
first_pos = self.start_pos
|
|
while True:
|
|
token_type, tok = self.next()
|
|
if tok != '.':
|
|
break
|
|
token_type, tok = self.next()
|
|
if token_type != tokenize.NAME:
|
|
break
|
|
append((tok, self.start_pos))
|
|
|
|
n = Name(names, first_pos, self.end_pos) if names else None
|
|
return (n, token_type, tok)
|
|
|
|
def _parseimportlist(self):
|
|
"""
|
|
The parser for the imports. Unlike the class and function parse
|
|
function, this returns no Import class, but rather an import list,
|
|
which is then added later on.
|
|
The reason, why this is not done in the same class lies in the nature
|
|
of imports. There are two ways to write them:
|
|
|
|
- from ... import ...
|
|
- import ...
|
|
|
|
To distinguish, this has to be processed after the parser.
|
|
|
|
:return: List of imports.
|
|
:rtype: list
|
|
"""
|
|
imports = []
|
|
brackets = False
|
|
continue_kw = [",", ";", "\n", ')'] \
|
|
+ list(set(keyword.kwlist) - set(['as']))
|
|
while True:
|
|
defunct = False
|
|
token_type, tok = self.next()
|
|
if token_type == tokenize.ENDMARKER:
|
|
break
|
|
if brackets and tok == '\n':
|
|
self.next()
|
|
if tok == '(': # python allows only one `(` in the statement.
|
|
brackets = True
|
|
self.next()
|
|
i, token_type, tok = self._parsedotname(self.current)
|
|
if not i:
|
|
defunct = True
|
|
name2 = None
|
|
if tok == 'as':
|
|
name2, token_type, tok = self._parsedotname()
|
|
imports.append((i, name2, defunct))
|
|
while tok not in continue_kw:
|
|
token_type, tok = self.next()
|
|
if not (tok == "," or brackets and tok == '\n'):
|
|
break
|
|
return imports
|
|
|
|
def _parseparen(self):
|
|
"""
|
|
Functions and Classes have params (which means for classes
|
|
super-classes). They are parsed here and returned as Statements.
|
|
|
|
:return: List of Statements
|
|
:rtype: list
|
|
"""
|
|
names = []
|
|
tok = None
|
|
pos = 0
|
|
breaks = [',', ':']
|
|
while tok not in [')', ':']:
|
|
param, tok = self._parse_statement(added_breaks=breaks,
|
|
stmt_class=Param)
|
|
if param and tok == ':':
|
|
# parse annotations
|
|
annotation, tok = self._parse_statement(added_breaks=breaks)
|
|
param.add_annotation(annotation)
|
|
|
|
if param:
|
|
param.position_nr = pos
|
|
names.append(param)
|
|
pos += 1
|
|
|
|
return names
|
|
|
|
def _parsefunction(self):
|
|
"""
|
|
The parser for a text functions. Process the tokens, which follow a
|
|
function definition.
|
|
|
|
:return: Return a Scope representation of the tokens.
|
|
:rtype: Function
|
|
"""
|
|
first_pos = self.start_pos
|
|
token_type, fname = self.next()
|
|
if token_type != tokenize.NAME:
|
|
return None
|
|
|
|
fname = Name([(fname, self.start_pos)], self.start_pos, self.end_pos)
|
|
|
|
token_type, open = self.next()
|
|
if open != '(':
|
|
return None
|
|
params = self._parseparen()
|
|
|
|
token_type, colon = self.next()
|
|
annotation = None
|
|
if colon in ['-', '->']:
|
|
# parse annotations
|
|
if colon == '-':
|
|
# The Python 2 tokenizer doesn't understand this
|
|
token_type, colon = self.next()
|
|
if colon != '>':
|
|
return None
|
|
annotation, colon = self._parse_statement(added_breaks=[':'])
|
|
|
|
if colon != ':':
|
|
return None
|
|
|
|
# because of 2 line func param definitions
|
|
scope = Function(fname, params, first_pos, annotation)
|
|
if self.user_scope and scope != self.user_scope \
|
|
and self.user_position > first_pos:
|
|
self.user_scope = scope
|
|
return scope
|
|
|
|
def _parseclass(self):
|
|
"""
|
|
The parser for a text class. Process the tokens, which follow a
|
|
class definition.
|
|
|
|
:return: Return a Scope representation of the tokens.
|
|
:rtype: Class
|
|
"""
|
|
first_pos = self.start_pos
|
|
token_type, cname = self.next()
|
|
if token_type != tokenize.NAME:
|
|
debug.warning("class: syntax err, token is not a name@%s (%s: %s)"
|
|
% (self.start_pos[0], tokenize.tok_name[token_type], cname))
|
|
return None
|
|
|
|
cname = Name([(cname, self.start_pos)], self.start_pos, self.end_pos)
|
|
|
|
super = []
|
|
token_type, next = self.next()
|
|
if next == '(':
|
|
super = self._parseparen()
|
|
token_type, next = self.next()
|
|
|
|
if next != ':':
|
|
debug.warning("class syntax: %s@%s" % (cname, self.start_pos[0]))
|
|
return None
|
|
|
|
# because of 2 line class initializations
|
|
scope = Class(cname, super, first_pos)
|
|
if self.user_scope and scope != self.user_scope \
|
|
and self.user_position > first_pos:
|
|
self.user_scope = scope
|
|
return scope
|
|
|
|
def _parse_statement(self, pre_used_token=None, added_breaks=None,
|
|
stmt_class=Statement, list_comp=False):
|
|
"""
|
|
Parses statements like:
|
|
|
|
>>> a = test(b)
|
|
>>> a += 3 - 2 or b
|
|
|
|
and so on. One line at a time.
|
|
|
|
:param pre_used_token: The pre parsed token.
|
|
:type pre_used_token: set
|
|
:return: Statement + last parsed token.
|
|
:rtype: (Statement, str)
|
|
"""
|
|
|
|
string = ''
|
|
set_vars = []
|
|
used_funcs = []
|
|
used_vars = []
|
|
level = 0 # The level of parentheses
|
|
is_return = None
|
|
|
|
if pre_used_token:
|
|
token_type, tok = pre_used_token
|
|
else:
|
|
token_type, tok = self.next()
|
|
|
|
first_pos = self.start_pos
|
|
opening_brackets = ['{', '(', '[']
|
|
closing_brackets = ['}', ')', ']']
|
|
|
|
# the difference between "break" and "always break" is that the latter
|
|
# will even break in parentheses. This is true for typical flow
|
|
# commands like def and class and the imports, which will never be used
|
|
# in a statement.
|
|
breaks = ['\n', ':', ')']
|
|
always_break = [';', 'import', 'from', 'class', 'def', 'try', 'except',
|
|
'finally', 'while']
|
|
if added_breaks:
|
|
breaks += added_breaks
|
|
|
|
tok_list = []
|
|
while not (tok in always_break or tok in breaks and level <= 0):
|
|
set_string = None
|
|
#print 'parse_stmt', tok, tokenize.tok_name[token_type]
|
|
tok_list.append(self.current + (self.start_pos,))
|
|
if tok == 'as':
|
|
string += " %s " % tok
|
|
token_type, tok = self.next()
|
|
if token_type == tokenize.NAME:
|
|
n, token_type, tok = self._parsedotname(self.current)
|
|
if n:
|
|
set_vars.append(n)
|
|
tok_list.append(n)
|
|
string += ".".join(n.names)
|
|
continue
|
|
elif token_type == tokenize.NAME:
|
|
if tok in ['return', 'yield', 'del', 'raise', 'assert']:
|
|
if len(tok_list) > 1:
|
|
# this happens, when a statement has opening brackets,
|
|
# which are not closed again, here I just start a new
|
|
# statement. This is a hack, but I could not come up
|
|
# with a better solution.
|
|
# This is basically a reset of the statement.
|
|
debug.warning('keyword in statement %s@%s', tok_list,
|
|
self.start_pos[0])
|
|
tok_list = [self.current + (self.start_pos,)]
|
|
set_vars = []
|
|
used_funcs = []
|
|
used_vars = []
|
|
level = 0
|
|
set_string = tok + ' '
|
|
if tok in ['return', 'yield']:
|
|
is_return = tok
|
|
elif tok == 'for':
|
|
# list comprehensions!
|
|
middle, tok = self._parse_statement(added_breaks=['in'])
|
|
if tok != 'in' or middle is None:
|
|
if middle is None:
|
|
level -= 1
|
|
debug.warning('list comprehension formatting @%s' %
|
|
self.start_pos[0])
|
|
continue
|
|
|
|
b = [')', ']']
|
|
in_clause, tok = self._parse_statement(added_breaks=b,
|
|
list_comp=True)
|
|
if tok not in b or in_clause is None:
|
|
if in_clause is None:
|
|
self.gen.push_back(self._current_full)
|
|
debug.warning('list comprehension in_clause %s@%s' %
|
|
(tok, self.start_pos[0]))
|
|
continue
|
|
other_level = 0
|
|
|
|
for i, tok in enumerate(reversed(tok_list)):
|
|
if not isinstance(tok, Name):
|
|
tok = tok[1]
|
|
if tok in closing_brackets:
|
|
other_level -= 1
|
|
elif tok in opening_brackets:
|
|
other_level += 1
|
|
if other_level > 0:
|
|
break
|
|
else:
|
|
i = 0 # could not detect brackets -> nested list comp
|
|
|
|
tok_list, toks = tok_list[:-i], tok_list[-i:-1]
|
|
src = ''
|
|
for t in toks:
|
|
src += t[1] if isinstance(t, tuple) else t.get_code()
|
|
st = Statement(src, [], [], [], \
|
|
toks, first_pos, self.end_pos)
|
|
|
|
for s in [st, middle, in_clause]:
|
|
s.parent = weakref.ref(self.scope)
|
|
tok = ListComprehension(st, middle, in_clause)
|
|
tok_list.append(tok)
|
|
if list_comp:
|
|
string = ''
|
|
string += tok.get_code()
|
|
continue
|
|
else:
|
|
n, token_type, tok = self._parsedotname(self.current)
|
|
tok_list.pop() # removed last entry, because we add Name
|
|
if n:
|
|
tok_list.append(n)
|
|
if tok == '(':
|
|
# it must be a function
|
|
used_funcs.append(n)
|
|
else:
|
|
used_vars.append(n)
|
|
if string and re.match(r'[\w\d\'"]', string[-1]):
|
|
string += ' '
|
|
string += ".".join(n.names)
|
|
continue
|
|
elif '=' in tok and not tok in ['>=', '<=', '==', '!=']:
|
|
# there has been an assignement -> change vars
|
|
if level == 0:
|
|
set_vars = used_vars
|
|
used_vars = []
|
|
elif tok in opening_brackets:
|
|
level += 1
|
|
elif tok in closing_brackets:
|
|
level -= 1
|
|
|
|
if set_string is not None:
|
|
string = set_string
|
|
else:
|
|
string += tok
|
|
token_type, tok = self.next()
|
|
if not string:
|
|
return None, tok
|
|
#print 'new_stat', string, set_vars, used_funcs, used_vars
|
|
if self.freshscope and not self.no_docstr and len(tok_list) == 1 \
|
|
and self.last_token[0] == tokenize.STRING:
|
|
self.scope.add_docstr(self.last_token[1])
|
|
return None, tok
|
|
else:
|
|
stmt = stmt_class(string, set_vars, used_funcs, used_vars, \
|
|
tok_list, first_pos, self.end_pos)
|
|
self._check_user_stmt(stmt)
|
|
if not isinstance(stmt, Param):
|
|
for tok_name in self.module.temp_used_names:
|
|
try:
|
|
self.module.used_names[tok_name].add(stmt)
|
|
except KeyError:
|
|
self.module.used_names[tok_name] = set([stmt])
|
|
self.module.temp_used_names = []
|
|
if is_return:
|
|
# add returns to the scope
|
|
func = self.scope.get_parent_until(Function)
|
|
if is_return == 'yield':
|
|
func.is_generator = True
|
|
try:
|
|
func.returns.append(stmt)
|
|
except AttributeError:
|
|
debug.warning('return in non-function')
|
|
|
|
if tok in always_break:
|
|
self.gen.push_back(self._current_full)
|
|
return stmt, tok
|
|
|
|
def next(self):
|
|
""" Generate the next tokenize pattern. """
|
|
self._current_full = next(self.gen)
|
|
type, tok, self._tokenize_start_pos, self._tokenize_end_pos, \
|
|
self.parserline = self._current_full
|
|
if self.user_position and (self.start_pos[0] == self.user_position[0]
|
|
or self.user_scope is None
|
|
and self.start_pos[0] >= self.user_position[0]):
|
|
debug.dbg('user scope found [%s] = %s' % \
|
|
(self.parserline.replace('\n', ''), repr(self.scope)))
|
|
self.user_scope = self.scope
|
|
self.last_token = self.current
|
|
self.current = (type, tok)
|
|
return self.current
|
|
|
|
def parse(self):
|
|
"""
|
|
The main part of the program. It analyzes the given code-text and
|
|
returns a tree-like scope. For a more detailed description, see the
|
|
class description.
|
|
|
|
:param text: The code which should be parsed.
|
|
:param type: str
|
|
|
|
:raises: IndentationError
|
|
"""
|
|
buf = BytesIO(self.code)
|
|
self.gen = PushBackIterator(tokenize_func(buf.readline))
|
|
|
|
extended_flow = ['else', 'elif', 'except', 'finally']
|
|
statement_toks = ['{', '[', '(', '`']
|
|
|
|
decorators = []
|
|
self.freshscope = True
|
|
while True:
|
|
try:
|
|
token_type, tok = self.next()
|
|
#debug.dbg('main: tok=[%s] type=[%s] indent=[%s]'\
|
|
# % (tok, tokenize.tok_name[token_type], start_position[0]))
|
|
|
|
while token_type == tokenize.DEDENT \
|
|
and self.scope != self.module:
|
|
token_type, tok = self.next()
|
|
if self.start_pos[1] <= self.scope.start_pos[1]:
|
|
self.scope.end_pos = self.start_pos
|
|
self.scope = self.scope.parent()
|
|
|
|
# check again for unindented stuff. this is true for syntax
|
|
# errors. only check for names, because thats relevant here. If
|
|
# some docstrings are not indented, I don't care.
|
|
while self.start_pos[1] <= self.scope.start_pos[1] \
|
|
and (token_type == tokenize.NAME or tok in ['(', '['])\
|
|
and self.scope != self.module:
|
|
self.scope.end_pos = self.start_pos
|
|
self.scope = self.scope.parent()
|
|
|
|
first_pos = self.start_pos
|
|
if tok == 'def':
|
|
func = self._parsefunction()
|
|
if func is None:
|
|
debug.warning("function: syntax error@%s" %
|
|
self.start_pos[0])
|
|
continue
|
|
self.freshscope = True
|
|
self.scope = self.scope.add_scope(func, decorators)
|
|
decorators = []
|
|
elif tok == 'class':
|
|
cls = self._parseclass()
|
|
if cls is None:
|
|
debug.warning("class: syntax error@%s" %
|
|
self.start_pos[0])
|
|
continue
|
|
self.freshscope = True
|
|
self.scope = self.scope.add_scope(cls, decorators)
|
|
decorators = []
|
|
# import stuff
|
|
elif tok == 'import':
|
|
imports = self._parseimportlist()
|
|
for m, alias, defunct in imports:
|
|
i = Import(first_pos, self.end_pos, m, alias,
|
|
defunct=defunct)
|
|
self._check_user_stmt(i)
|
|
self.scope.add_import(i)
|
|
if not imports:
|
|
i = Import(first_pos, self.end_pos, None, defunct=True)
|
|
self._check_user_stmt(i)
|
|
self.freshscope = False
|
|
elif tok == 'from':
|
|
defunct = False
|
|
# take care for relative imports
|
|
relative_count = 0
|
|
while 1:
|
|
token_type, tok = self.next()
|
|
if tok != '.':
|
|
break
|
|
relative_count += 1
|
|
# the from import
|
|
mod, token_type, tok = self._parsedotname(self.current)
|
|
if str(mod) == 'import' and relative_count:
|
|
self.gen.push_back(self._current_full)
|
|
tok = 'import'
|
|
mod = None
|
|
if not mod and not relative_count or tok != "import":
|
|
debug.warning("from: syntax error@%s" %
|
|
self.start_pos[0])
|
|
defunct = True
|
|
if tok != 'import':
|
|
self.gen.push_back(self._current_full)
|
|
names = self._parseimportlist()
|
|
for name, alias, defunct2 in names:
|
|
star = name is not None and name.names[0] == '*'
|
|
if star:
|
|
name = None
|
|
i = Import(first_pos, self.end_pos, name, alias, mod,
|
|
star, relative_count, defunct=defunct or defunct2)
|
|
self._check_user_stmt(i)
|
|
self.scope.add_import(i)
|
|
self.freshscope = False
|
|
#loops
|
|
elif tok == 'for':
|
|
set_stmt, tok = self._parse_statement(added_breaks=['in'])
|
|
if tok == 'in':
|
|
statement, tok = self._parse_statement()
|
|
if tok == ':':
|
|
f = ForFlow([statement], first_pos,
|
|
set_stmt)
|
|
self.scope = self.scope.add_statement(f)
|
|
|
|
elif tok in ['if', 'while', 'try', 'with'] + extended_flow:
|
|
added_breaks = []
|
|
command = tok
|
|
if command in ['except', 'with']:
|
|
added_breaks.append(',')
|
|
# multiple statements because of with
|
|
inits = []
|
|
first = True
|
|
while first or command == 'with' and tok != ':':
|
|
statement, tok = \
|
|
self._parse_statement(added_breaks=added_breaks)
|
|
if command == 'except' and tok in added_breaks:
|
|
# the except statement defines a var
|
|
# this is only true for python 2
|
|
n, token_type, tok = self._parsedotname()
|
|
if n:
|
|
statement.set_vars.append(n)
|
|
statement.code += ',' + n.get_code()
|
|
if statement:
|
|
inits.append(statement)
|
|
first = False
|
|
|
|
if tok == ':':
|
|
f = Flow(command, inits, first_pos)
|
|
if command in extended_flow:
|
|
# the last statement has to be another part of
|
|
# the flow statement, because a dedent releases the
|
|
# main scope, so just take the last statement.
|
|
try:
|
|
s = self.scope.statements[-1].set_next(f)
|
|
except (AttributeError, IndexError):
|
|
# If set_next doesn't exist, just add it.
|
|
s = self.scope.add_statement(f)
|
|
else:
|
|
s = self.scope.add_statement(f)
|
|
self.scope = s
|
|
else:
|
|
debug.warning('syntax err, flow started @%s',
|
|
self.start_pos[0])
|
|
# globals
|
|
elif tok == 'global':
|
|
stmt, tok = self._parse_statement(self.current)
|
|
if stmt:
|
|
self.scope.add_statement(stmt)
|
|
for name in stmt.used_vars:
|
|
# add the global to the top, because there it is
|
|
# important.
|
|
self.module.add_global(name)
|
|
# decorator
|
|
elif tok == '@':
|
|
stmt, tok = self._parse_statement()
|
|
decorators.append(stmt)
|
|
elif tok == 'pass':
|
|
continue
|
|
# default
|
|
elif token_type in [tokenize.NAME, tokenize.STRING,
|
|
tokenize.NUMBER] \
|
|
or tok in statement_toks:
|
|
# this is the main part - a name can be a function or a
|
|
# normal var, which can follow anything. but this is done
|
|
# by the statement parser.
|
|
stmt, tok = self._parse_statement(self.current)
|
|
if stmt:
|
|
self.scope.add_statement(stmt)
|
|
self.freshscope = False
|
|
else:
|
|
if token_type not in [tokenize.COMMENT, tokenize.INDENT,
|
|
tokenize.NEWLINE, tokenize.NL,
|
|
tokenize.ENDMARKER]:
|
|
debug.warning('token not classified', tok, token_type,
|
|
self.start_pos[0])
|
|
except StopIteration: # thrown on EOF
|
|
break
|
|
except tokenize.TokenError:
|
|
# We just ignore this error, I try to handle it earlier - as
|
|
# good as possible
|
|
debug.warning('parentheses not closed error')
|
|
except IndentationError:
|
|
# This is an error, that tokenize may produce, because the code
|
|
# is not indented as it should. Here it just ignores this line
|
|
# and restarts the parser.
|
|
# (This is a rather unlikely error message, for normal code,
|
|
# tokenize seems to be pretty tolerant)
|
|
self._line_of_tokenize_restart = self.start_pos[0] + 1
|
|
self._tokenize_start_pos = (0, 0)
|
|
self._tokenize_end_pos = (0, 0)
|
|
debug.warning('indentation error on line %s, ignoring it' %
|
|
(self.start_pos[0]))
|
|
self.gen = PushBackIterator(tokenize_func(buf.readline))
|
|
return self.module
|