remove token.py, docstrings are now just normal tokens until used

jedi/api/classes.py

@@ -10,6 +10,7 @@ from jedi import settings
 from jedi import common
 from jedi import cache
 from jedi.parser import representation as pr
+from jedi.parser.tokenize import Token
 from jedi.evaluate import representation as er
 from jedi.evaluate import iterable
 from jedi.evaluate import imports
@@ -241,8 +242,10 @@ class BaseDefinition(object):
 
         See :attr:`doc` for example.
         """
-        if isinstance(self._definition.docstr, pr.token_pr.TokenDocstring):
+        print self._definition.docstr
-            return unicode(self._definition.docstr.as_string())
+        if isinstance(self._definition.docstr, Token):
+# TODO again ugly, we should have direct access.
+            return unicode(self._definition.docstr.string)
         try:
             return unicode(self._definition.docstr)
         except AttributeError:

jedi/evaluate/compiled/fake.py

@@ -9,7 +9,7 @@ import inspect
 
 from jedi._compatibility import is_py3, builtins, unicode
 from jedi.parser import Parser
-from jedi.parser import token as token_pr
+from jedi.parser import tokenize
 from jedi.parser.representation import Class
 from jedi.evaluate.helpers import FakeName
 
@@ -105,7 +105,8 @@ def get_faked(module, obj, name=None):
         # contain it).
         result.docstr = None
         if obj.__doc__:
-            result.docstr = token_pr.TokenDocstring.fake_docstring(obj.__doc__)
+            doc = '''"""%s"""''' % obj.__doc__  # TODO need escapes.
+            result.docstr = tokenize.Token(tokenize.STRING, doc, (0, 0))
         return result
 
 

jedi/evaluate/docstrings.py

@@ -18,6 +18,7 @@ import re
 
 from jedi.evaluate.cache import memoize_default
 from jedi.parser import Parser
+from jedi.parser.representation import docstring_content
 
 DOCSTRING_PARAM_PATTERNS = [
     r'\s*:type\s+%s:\s*([^\n]+)',  # Sphinx
@@ -39,7 +40,8 @@ def follow_param(evaluator, param):
     if not func.docstr:
         return []
     param_str = _search_param_in_docstr(
-        func.docstr.as_string(),
+# TODO  this is ugly, no direct access?
+        docstring_content(func.docstr),
         str(param.get_name())
     )
     position = (1, 0)
@@ -119,7 +121,7 @@ def find_return_types(evaluator, func):
 
     if not func.docstr:
         return []
-    type_str = search_return_in_docstr(func.docstr.as_string())
+    type_str = search_return_in_docstr(docstring_content(func.docstr))
     if not type_str:
         return []
 

jedi/parser/__init__.py

@@ -21,7 +21,6 @@ from jedi._compatibility import next
 from jedi import debug
 from jedi import common
 from jedi.parser import representation as pr
-from jedi.parser import token as token_pr
 from jedi.parser import tokenize
 
 
@@ -358,9 +357,7 @@ class Parser(object):
                 and first_tok.type == tokenize.STRING:
             # Normal docstring check
             if self.freshscope and not self.no_docstr:
-                self._scope.add_docstr(
+                self._scope.add_docstr(first_tok)
-                    token_pr.TokenDocstring(first_tok)
-                )
                 return None, tok
 
             # Attribute docstring (PEP 224) support (sphinx uses it, e.g.)
@@ -368,9 +365,7 @@ class Parser(object):
             elif first_tok.type == tokenize.STRING:
                 with common.ignored(IndexError, AttributeError):
                     # ...then set it as a docstring
-                    self._scope.statements[-1].add_docstr(
+                    self._scope.statements[-1].add_docstr(first_tok)
-                        token_pr.TokenDocstring(first_tok)
-                    )
                     return None, tok
 
         stmt = stmt_class(self.module, tok_list, first_pos, tok.end_pos,

jedi/parser/representation.py

@@ -44,12 +44,16 @@ from jedi import common
 from jedi import debug
 from jedi import cache
 from jedi.parser import tokenize
-from jedi.parser import token as token_pr
 
 
 SCOPE_CONTENTS = ['asserts', 'subscopes', 'imports', 'statements', 'returns']
 
 
+def docstring_content(token):
+    """Returns a literal cleaned version of the ``Token``."""
+    return unicode(cleandoc(literal_eval(token.string)))
+
+
 class GetCodeState(object):
     """A helper class for passing the state of get_code in a thread-safe
     manner"""
@@ -241,7 +245,7 @@ class Scope(Simple, IsScope):
         """
         string = ""
         if self.docstr:
-            string += '"""' + self.docstr.as_string() + '"""\n'
+            string += '"""' + docstring_content(self.docstr) + '"""\n'
 
         objs = self.subscopes + self.imports + self.statements + self.returns
         for obj in sorted(objs, key=lambda x: x.start_pos):
@@ -481,7 +485,7 @@ class Class(Scope):
         """
         docstr = ""
         if self.docstr:
-            docstr = self.docstr.as_string()
+            docstr = docstring_content(self.docstr)
         for sub in self.subscopes:
             if sub.name.names[-1] == '__init__':
                 return '%s\n\n%s' % (
@@ -571,7 +575,7 @@ class Function(Scope):
         """ Return a document string including call signature. """
         docstr = ""
         if self.docstr:
-            docstr = self.docstr.as_string()
+            docstr = docstring_content(self.docstr)
         return '%s\n\n%s' % (
             self.get_call_signature(),
             docstr,
@@ -849,7 +853,7 @@ isinstance(c, tokenize.Token) else unicode(c)
         code = ''.join(assemble(*a) for a in self.assignment_details)
         code += assemble(self.expression_list())
         if self.docstr:
-            code += '\n"""%s"""' % self.docstr.as_string()
+            code += '\n"""%s"""' % docstring_content(self.docstr)
 
         if new_line:
             return code + '\n'

jedi/parser/token.py

@@ -1,35 +0,0 @@
-# -*- coding: utf-8 -*-
-""" Efficient representation of tokens
-
-We want to have a token_list and start_position for everything the
-tokenizer returns. Therefore we need a memory efficient class. We
-found that a flat object with slots is the best.
-"""
-from inspect import cleandoc
-from ast import literal_eval
-
-from jedi._compatibility import unicode
-from jedi.parser.tokenize import Token
-
-
-class TokenDocstring(Token):
-    """A string token that is a docstring.
-
-    as_string() will clean the token representing the docstring.
-    """
-    __slots__ = ()
-
-    def __init__(self, token=None, state=None):
-        if token:
-            self.__setstate__(token.__getstate__())
-        else:
-            self.__setstate__(state)
-
-    @classmethod
-    def fake_docstring(cls, docstr):
-        # TODO: fixme when tests are up again
-        return TokenDocstring(state=(0, '"""\n%s\n"""' % docstr, 0, 0))
-
-    def as_string(self):
-        """Returns a literal cleaned version of the token"""
-        return unicode(cleandoc(literal_eval(self.string)))