##// END OF EJS Templates
Reset the interactive namespace __warningregistry__ before executing code...
Reset the interactive namespace __warningregistry__ before executing code Fixes #6611. Idea: Right now, people often don't see important warnings when running code in IPython, because (to a first approximation) any given warning will only issue once per session. Blink and you'll miss it! This is a very common contributor to confused emails to numpy-discussion. E.g.: In [5]: 1 / my_array_with_random_contents /home/njs/.user-python2.7-64bit-3/bin/ipython:1: RuntimeWarning: divide by zero encountered in divide #!/home/njs/.user-python2.7-64bit-3/bin/python Out[5]: array([ 1.77073316, -2.29765021, -2.01800811, ..., 1.13871243, -1.08302964, -8.6185091 ]) Oo, right, guess I gotta be careful of those zeros -- thanks, numpy, for giving me that warning! A few days later: In [592]: 1 / some_other_array Out[592]: array([ 3.07735763, 0.50769289, 0.83984078, ..., -0.67563917, -0.85736257, -1.36511271]) Oops, it turns out that this array had a zero in it too, and that's going to bite me later. But no warning this time! The effect of this commit is to make it so that warnings triggered by the code in cell 5 do *not* suppress warnings triggered by the code in cell 592. Note that this only applies to warnings triggered *directly* by code entered interactively -- if somepkg.foo() calls anotherpkg.bad_func() which issues a warning, then this warning will still only be displayed once, even if multiple cells call somepkg.foo(). But if cell 5 and cell 592 both call anotherpkg.bad_func() directly, then both will get warnings. (Important exception: if foo() is defined *interactively*, and calls anotherpkg.bad_func(), then every cell that calls foo() will display the warning again. This is unavoidable without fixes to CPython upstream.) Explanation: Python's warning system has some weird quirks. By default, it tries to suppress duplicate warnings, where "duplicate" means the same warning message triggered twice by the same line of code. This requires determining which line of code is responsible for triggering a warning, and this is controlled by the stacklevel= argument to warnings.warn. Basically, though, the idea is that if foo() calls bar() which calls baz() which calls some_deprecated_api(), then baz() will get counted as being "responsible", and the warning system will make a note that the usage of some_deprecated_api() inside baz() has already been warned about and doesn't need to be warned about again. So far so good. To accomplish this, obviously, there has to be a record of somewhere which line this was. You might think that this would be done by recording the filename:linenumber pair in a dict inside the warnings module, or something like that. You would be wrong. What actually happens is that the warnings module will use stack introspection to reach into baz()'s execution environment, create a global (module-level) variable there named __warningregistry__, and then, inside this dictionary, record just the line number. Basically, it assumes that any given module contains only one line 1, only one line 2, etc., so storing the filename is irrelevant. Obviously for interactive code this is totally wrong -- all cells share the same execution environment and global namespace, and they all contain a new line 1. Currently the warnings module treats these as if they were all the same line. In fact they are not the same line; once we have executed a given chunk of code, we will never see those particular lines again. As soon as a given chunk of code finishes executing, its line number labels become meaningless, and the corresponding warning registry entries become meaningless as well. Therefore, with this patch we delete the __warningregistry__ each time we execute a new block of code.

File last commit:

r13392:bace1f46
r18548:61431d7d
Show More
ipstruct.py
391 lines | 11.6 KiB | text/x-python | PythonLexer
# encoding: utf-8
"""A dict subclass that supports attribute style access.
Authors:
* Fernando Perez (original)
* Brian Granger (refactoring to a dict subclass)
"""
#-----------------------------------------------------------------------------
# Copyright (C) 2008-2011 The IPython Development Team
#
# Distributed under the terms of the BSD License. The full license is in
# the file COPYING, distributed as part of this software.
#-----------------------------------------------------------------------------
#-----------------------------------------------------------------------------
# Imports
#-----------------------------------------------------------------------------
__all__ = ['Struct']
#-----------------------------------------------------------------------------
# Code
#-----------------------------------------------------------------------------
class Struct(dict):
"""A dict subclass with attribute style access.
This dict subclass has a a few extra features:
* Attribute style access.
* Protection of class members (like keys, items) when using attribute
style access.
* The ability to restrict assignment to only existing keys.
* Intelligent merging.
* Overloaded operators.
"""
_allownew = True
def __init__(self, *args, **kw):
"""Initialize with a dictionary, another Struct, or data.
Parameters
----------
args : dict, Struct
Initialize with one dict or Struct
kw : dict
Initialize with key, value pairs.
Examples
--------
>>> s = Struct(a=10,b=30)
>>> s.a
10
>>> s.b
30
>>> s2 = Struct(s,c=30)
>>> sorted(s2.keys())
['a', 'b', 'c']
"""
object.__setattr__(self, '_allownew', True)
dict.__init__(self, *args, **kw)
def __setitem__(self, key, value):
"""Set an item with check for allownew.
Examples
--------
>>> s = Struct()
>>> s['a'] = 10
>>> s.allow_new_attr(False)
>>> s['a'] = 10
>>> s['a']
10
>>> try:
... s['b'] = 20
... except KeyError:
... print('this is not allowed')
...
this is not allowed
"""
if not self._allownew and key not in self:
raise KeyError(
"can't create new attribute %s when allow_new_attr(False)" % key)
dict.__setitem__(self, key, value)
def __setattr__(self, key, value):
"""Set an attr with protection of class members.
This calls :meth:`self.__setitem__` but convert :exc:`KeyError` to
:exc:`AttributeError`.
Examples
--------
>>> s = Struct()
>>> s.a = 10
>>> s.a
10
>>> try:
... s.get = 10
... except AttributeError:
... print("you can't set a class member")
...
you can't set a class member
"""
# If key is an str it might be a class member or instance var
if isinstance(key, str):
# I can't simply call hasattr here because it calls getattr, which
# calls self.__getattr__, which returns True for keys in
# self._data. But I only want keys in the class and in
# self.__dict__
if key in self.__dict__ or hasattr(Struct, key):
raise AttributeError(
'attr %s is a protected member of class Struct.' % key
)
try:
self.__setitem__(key, value)
except KeyError as e:
raise AttributeError(e)
def __getattr__(self, key):
"""Get an attr by calling :meth:`dict.__getitem__`.
Like :meth:`__setattr__`, this method converts :exc:`KeyError` to
:exc:`AttributeError`.
Examples
--------
>>> s = Struct(a=10)
>>> s.a
10
>>> type(s.get)
<... 'builtin_function_or_method'>
>>> try:
... s.b
... except AttributeError:
... print("I don't have that key")
...
I don't have that key
"""
try:
result = self[key]
except KeyError:
raise AttributeError(key)
else:
return result
def __iadd__(self, other):
"""s += s2 is a shorthand for s.merge(s2).
Examples
--------
>>> s = Struct(a=10,b=30)
>>> s2 = Struct(a=20,c=40)
>>> s += s2
>>> sorted(s.keys())
['a', 'b', 'c']
"""
self.merge(other)
return self
def __add__(self,other):
"""s + s2 -> New Struct made from s.merge(s2).
Examples
--------
>>> s1 = Struct(a=10,b=30)
>>> s2 = Struct(a=20,c=40)
>>> s = s1 + s2
>>> sorted(s.keys())
['a', 'b', 'c']
"""
sout = self.copy()
sout.merge(other)
return sout
def __sub__(self,other):
"""s1 - s2 -> remove keys in s2 from s1.
Examples
--------
>>> s1 = Struct(a=10,b=30)
>>> s2 = Struct(a=40)
>>> s = s1 - s2
>>> s
{'b': 30}
"""
sout = self.copy()
sout -= other
return sout
def __isub__(self,other):
"""Inplace remove keys from self that are in other.
Examples
--------
>>> s1 = Struct(a=10,b=30)
>>> s2 = Struct(a=40)
>>> s1 -= s2
>>> s1
{'b': 30}
"""
for k in other.keys():
if k in self:
del self[k]
return self
def __dict_invert(self, data):
"""Helper function for merge.
Takes a dictionary whose values are lists and returns a dict with
the elements of each list as keys and the original keys as values.
"""
outdict = {}
for k,lst in data.items():
if isinstance(lst, str):
lst = lst.split()
for entry in lst:
outdict[entry] = k
return outdict
def dict(self):
return self
def copy(self):
"""Return a copy as a Struct.
Examples
--------
>>> s = Struct(a=10,b=30)
>>> s2 = s.copy()
>>> type(s2) is Struct
True
"""
return Struct(dict.copy(self))
def hasattr(self, key):
"""hasattr function available as a method.
Implemented like has_key.
Examples
--------
>>> s = Struct(a=10)
>>> s.hasattr('a')
True
>>> s.hasattr('b')
False
>>> s.hasattr('get')
False
"""
return key in self
def allow_new_attr(self, allow = True):
"""Set whether new attributes can be created in this Struct.
This can be used to catch typos by verifying that the attribute user
tries to change already exists in this Struct.
"""
object.__setattr__(self, '_allownew', allow)
def merge(self, __loc_data__=None, __conflict_solve=None, **kw):
"""Merge two Structs with customizable conflict resolution.
This is similar to :meth:`update`, but much more flexible. First, a
dict is made from data+key=value pairs. When merging this dict with
the Struct S, the optional dictionary 'conflict' is used to decide
what to do.
If conflict is not given, the default behavior is to preserve any keys
with their current value (the opposite of the :meth:`update` method's
behavior).
Parameters
----------
__loc_data : dict, Struct
The data to merge into self
__conflict_solve : dict
The conflict policy dict. The keys are binary functions used to
resolve the conflict and the values are lists of strings naming
the keys the conflict resolution function applies to. Instead of
a list of strings a space separated string can be used, like
'a b c'.
kw : dict
Additional key, value pairs to merge in
Notes
-----
The `__conflict_solve` dict is a dictionary of binary functions which will be used to
solve key conflicts. Here is an example::
__conflict_solve = dict(
func1=['a','b','c'],
func2=['d','e']
)
In this case, the function :func:`func1` will be used to resolve
keys 'a', 'b' and 'c' and the function :func:`func2` will be used for
keys 'd' and 'e'. This could also be written as::
__conflict_solve = dict(func1='a b c',func2='d e')
These functions will be called for each key they apply to with the
form::
func1(self['a'], other['a'])
The return value is used as the final merged value.
As a convenience, merge() provides five (the most commonly needed)
pre-defined policies: preserve, update, add, add_flip and add_s. The
easiest explanation is their implementation::
preserve = lambda old,new: old
update = lambda old,new: new
add = lambda old,new: old + new
add_flip = lambda old,new: new + old # note change of order!
add_s = lambda old,new: old + ' ' + new # only for str!
You can use those four words (as strings) as keys instead
of defining them as functions, and the merge method will substitute
the appropriate functions for you.
For more complicated conflict resolution policies, you still need to
construct your own functions.
Examples
--------
This show the default policy:
>>> s = Struct(a=10,b=30)
>>> s2 = Struct(a=20,c=40)
>>> s.merge(s2)
>>> sorted(s.items())
[('a', 10), ('b', 30), ('c', 40)]
Now, show how to specify a conflict dict:
>>> s = Struct(a=10,b=30)
>>> s2 = Struct(a=20,b=40)
>>> conflict = {'update':'a','add':'b'}
>>> s.merge(s2,conflict)
>>> sorted(s.items())
[('a', 20), ('b', 70)]
"""
data_dict = dict(__loc_data__,**kw)
# policies for conflict resolution: two argument functions which return
# the value that will go in the new struct
preserve = lambda old,new: old
update = lambda old,new: new
add = lambda old,new: old + new
add_flip = lambda old,new: new + old # note change of order!
add_s = lambda old,new: old + ' ' + new
# default policy is to keep current keys when there's a conflict
conflict_solve = dict.fromkeys(self, preserve)
# the conflict_solve dictionary is given by the user 'inverted': we
# need a name-function mapping, it comes as a function -> names
# dict. Make a local copy (b/c we'll make changes), replace user
# strings for the three builtin policies and invert it.
if __conflict_solve:
inv_conflict_solve_user = __conflict_solve.copy()
for name, func in [('preserve',preserve), ('update',update),
('add',add), ('add_flip',add_flip),
('add_s',add_s)]:
if name in inv_conflict_solve_user.keys():
inv_conflict_solve_user[func] = inv_conflict_solve_user[name]
del inv_conflict_solve_user[name]
conflict_solve.update(self.__dict_invert(inv_conflict_solve_user))
for key in data_dict:
if key not in self:
self[key] = data_dict[key]
else:
self[key] = conflict_solve[key](self[key],data_dict[key])