Viewing file:  Namespace.py (12.78 KB) -rw-r--r--
Select action/file-type:
 (+) |  (+) |  (+) | Code (+) | Session (+) |  (+) | SDB (+) |  (+) |  (+) |  (+) |  (+) |  (+) |
""" Namespace objects.
Copyright (c) 1998-2000, Marc-Andre Lemburg; mailto:mal@lemburg.com
Copyright (c) 2000-2001, eGenix.com Software GmbH; mailto:info@egenix.com
See the documentation for further information on copyrights,
or contact the author. All Rights Reserved.
"""
import string,types
import mx.Tools.NewBuiltins
from mx.Tools import freeze
class Namespace:
""" Base class for Namespaces.
Instances of this namespace can store arbitrary attributes,
which are stored in the instances dictionary. They are then
available via attribute lookup, e.g. ns.attr, and index
lookup, e.g. ns['attr'].
The Namespaces support acquisition. To link the object to
another object, set self.baseobj to the other
object. Attributes not found in the Namespace instance are
then looked up in the other object as well.
NOTE: All methods and internal attributes must start with an
underscore. Namespace attributes must *not* start with an
underscore to prevent naming collisions with internal
variables and methods. Note that names with leading
underscore cannot be acquired.
"""
def __init__(self,**kws):
""" Initializes the namespace with the attributes given
as keywords.
"""
if kws:
self.__dict__.update(kws)
def _clear(self,save=(),recursive=0,
getattr=getattr):
""" Clears the namespace.
The attributes named in save are not cleared. Per default
all attributes are cleared. You may want to save the
baseobj attribute in case you use it.
Class variables are not cleared this way, but overriding
instance variables do get deleted.
If recursive is true (default is false), the namespace is
cleared recursively, meaning that all Namespace instances
contained in it are also cleared.
"""
d = {}
for attr in save:
d[attr] = getattr(self,attr)
# Clear the old instance dictionary (recursively) so that it
# can be garbage collected
if recursive:
for obj in self.__dict__.items():
if isinstance(obj, Namespace):
obj._clear()
obj = None
self.__dict__.clear()
# Assing the new one
self.__dict__ = d
def _update(self, other_namespace,
DictType=types.DictType,type=type):
""" Updates the namespace dictionary with the new entries
taken from the object other_namespace.
other_namespace can either be a dictionary or another
Namespace object.
"""
if type(other_namespace) is DictType:
self.__dict__.update(other_namespace)
else:
self.__dict__.update(other_namespace.__dict__)
def _load(self, other_namespace, save=(),
ignore_types=(types.BuiltinMethodType,
types.ClassType,
types.MethodType,
types.FunctionType),
DictType=types.DictType,type=type):
""" Updates the namespace with the updated entries taken from
the object other_namespace provided that the entries already
exist in the namespace dictionary
other_namespace can either be a dictionary or another
Namespace object.
Attributes listed in save are not overwritten. New values
having one of the types listed in ignore_types are also
ignored.
"""
if type(other_namespace) is DictType:
items = other_namespace.items()
else:
items = other_namespace._items()
d = self.__dict__
exists = d.has_key
for k, v in items:
if exists(k):
if k in save or \
type(v) in ignore_types:
continue
d[k] = v
exists = None
d = None
def _dictionary(self):
""" Returns the current namespace contents as dictionary.
The dictionaries values should be considered read-only in
case they are mutable.
"""
return self.__dict__.copy()
def _copy(self):
""" Returns a shallow copy of the current namespace.
See the standard Python module "copy" for details on the
copying mechanism.
"""
import copy
return copy.copy(self)
def _deepcopy(self):
""" Returns a deep copy of the current namespace.
See the standard Python module "copy" for details on the
copying mechanism.
"""
import copy
return copy.deepcopy(self)
### Index interface
def __getitem__(self,name,
getattr=getattr):
""" Index interface to the namespace.
Raises IndexErrors in case an attribute is not found.
"""
try:
return getattr(self, name)
except AttributeError,why:
raise IndexError,why
def __setitem__(self,name,value,
setattr=setattr):
""" Index interface to the namespace.
"""
return setattr(self,name,value)
def __len__(self):
""" Return the number of instance attributes stored in the
namespace.
"""
return len(self.__dict__)
def __nonzero__(self):
""" Returns 1/0 depending on whether any instance attributes
are defined in the namespace.
An empty Namespace is false, if it contains at least one
entry it is true.
"""
return (len(self.__dict__) > 0)
### Dictionary like interface to the instance dictionary
def _items(self):
""" Returns a list of tuples (attrname, attrvalue) for
all (instance) attributes stored in the object.
"""
return self.__dict__.items()
def _keys(self):
""" Returns a list of attribute names for
all (instance) attributes stored in the object.
"""
return self.__dict__.keys()
def _values(self):
""" Returns a list of attribute values for
all (instance) attributes stored in the object.
"""
return self.__dict__.values()
def _has_key(self, key):
""" Returns 1/0 depending on whether the object has
an instance attribute key or not.
Note that class attributes cannot be queried this way.
"""
return self.__dict__.has_key(key)
# Alias
_hasattr = _has_key
def _get(self,key,default=None):
""" Returns the value of attribute key, or default if no
such attribute is defined.
"""
return self.__dict__.get(key,default)
# Alias
_getattr = _get
### Additional methods:
def _extract(self,names,defaults=(),
extract=extract):
""" Extract a list of named objects from the namespace.
defaults may be given as sequence of the same size as
names to provide default values in case names are not
found in the namespace.
Only the instance dictionary is scanned, the class
variables are not available through this mechanism.
"""
return extract(self.__dict__,names,defaults)
def _prefixgroup(self,prefix,nonzero=0,wholename=0):
""" Find all attributes having prefix as common prefix and return
a dictionary mapping the attribute suffixes (the prefix
is removed) to their values.
If nonzero is set, then only fields with non-zero value
are included in the dictionary
If wholename is true, the prefix is not removed from the
attribute names.
"""
plen = len(prefix)
d = {}
if not nonzero and not wholename:
# Shortcut for the default settings
for k,v in self.__dict__.items():
if k[:plen] == prefix:
d[k[plen:]] = v
else:
for k,v in self.__dict__.items():
if k[:plen] == prefix:
if nonzero and not v:
continue
if wholename:
d[k] = v
else:
d[k[plen:]] = v
return d
### Acquistion:
baseobj = None # Base object ("containing" this object)
# Implicit acquisition
__getattr__ = acquire # Use the (new) builtin acquire() to implement
# the acquisition mechanism
### For debugging purposes:
def __repr__(self,
join=string.join,id=id):
keys = self.__dict__.keys()
keys.sort()
return '<%s.%s (%s) at 0x%x>' % (
self.__class__.__module__,
self.__class__.__name__,
join(keys,','),
id(self))
def __str__(self,
join=string.join,id=id):
items = self.__dict__.items()
items.sort()
l = ['%s.%s at 0x%x:' % (self.__class__.__module__,
self.__class__.__name__,
id(self))]
for k,v in items:
try:
s = repr(v)
except (ValueError, TypeError, AttributeError):
s = '*** repr(value) failed ***'
else:
if len(s) > 200:
s = s[:200] + '... (truncated, full length = %i)' % len(s)
l.append(' %-10s %s' % (k + ':', s))
return string.join(l,'\n')
###
class DefaultMixin:
""" Mixin that provides default attribute values.
The namespace returns _default for all undefined attributes
(except the ones that start with an underscore, see
__getattr__). The _default is defined as '' for this class.
The acquisition feature is *not* available for instances of this
class.
"""
_default = ''
def __getattr__(self,name):
""" Return self._default for all undefined attributes that don't
start with an underscore.
For attributes that start with an underscore, raise an
AttributeError. This is especially important if the
objects are to be pickled, since the pickle mechanism
queries a few methods on the objects, which it then calls.
"""
if name[0] == '_':
raise AttributeError,name
return self._default
###
class NamespaceWithDefault(DefaultMixin,Namespace):
""" Namespace class with an infinite number of attributes.
The namespace returns _default for all undefined attributes
(except the ones that start with an underscore, see
__getattr__). The _default is defined as '' for this class.
The acquisition feature is *not* available for instances of this
class.
"""
freeze(NamespaceWithDefault)
###
class _NotFound:
pass
_NotFound = _NotFound()
class FormatNamespaceMixin:
""" This mixin makes a Namespace class conform to the Python
format string mechanism.
The following format specifiers are supported:
%(attribute)s - gives the string representation
of .attribute
%(<eval-string>)s - evaluates the <eval-string> in the
Namespace dictionary and returns
the string representation
All standard Python string formatting output types are
supported, provided that the attribute or evaluation result
can be converted to these types.
"""
def __getitem__(self, item,
_NotFound=_NotFound):
obj = getattr(self, item, _NotFound)
if obj is not _NotFound:
return obj
return eval(item, self.__dict__)
###
class FormatNamespace(FormatNamespaceMixin,
Namespace):
""" Namespace class which supports the Python format string
lookup mechanism.
These Namespaces can be used in 'format string' % namespace
formatting.
The following format specifiers are supported:
%(attribute)s - gives the string representation
of .attribute
%(<eval-string>)s - evaluates the <eval-string> in the
Namespace dictionary and returns
the string representation
All standard Python string formatting output types are
supported, provided that the attribute or evaluation result
can be converted to these types.
"""
pass
freeze(FormatNamespace)
|