def __setitem__(self, key, value):
from corpkit.process import makesafe, is_number
if key.startswith('c') and len(key) > 1 and all(
is_number(x) for x in key[1:]):
self.__setattr__(key.lstrip('c'), value)
else:
self.__setattr__(key, value)
def makesafe(variabletext):
import re
from corpkit.process import is_number
variable_safe_r = re.compile('[\W_]+', re.UNICODE)
try:
txt = variabletext.name.split('.')[0].replace('-parsed', '')
except AttributeError:
txt = variabletext.split('.')[0].replace('-parsed', '')
variable_safe = re.sub(variable_safe_r, '', txt)
if is_number(variable_safe):
variable_safe = 'c' + variable_safe
return variable_safe
def __getitem__(self, key):
"""allow slicing, indexing"""
from corpkit.process import makesafe
if isinstance( key, slice ) :
#Get the start, stop, and step from the slice
return Corpora([self[ii] for ii in xrange(*key.indices(len(self)))])
elif type(key) == int:
return self.__getitem__(makesafe(self.data[key]))
else:
try:
return self.__getattribute__(key)
except:
from corpkit.process import is_number
if is_number(key):
return self.__getattribute__('c' + key)
def __getitem__(self, key):
from corpkit.process import makesafe
if isinstance(key, slice):
# Get the start, stop, and step from the slice
return Datalist([self[ii] for ii in range(
*key.indices(len(self.subcorpora)))])
elif isinstance(key, int):
return self.subcorpora.__getitem__(makesafe(self.subcorpora[key]))
else:
try:
return self.subcorpora.__getattribute__(key)
except:
from corpkit.process import is_number
if is_number(key):
return self.__getattribute__('c' + key)
def makesafe(variabletext, drop_datatype=True, hyphens_ok=False):
import re
from corpkit.process import is_number
if hyphens_ok:
variable_safe_r = re.compile(r'[^A-Za-z0-9_-]+', re.UNICODE)
else:
variable_safe_r = re.compile(r'[^A-Za-z0-9_]+', re.UNICODE)
try:
txt = variabletext.name.split('.')[0]
except AttributeError:
txt = variabletext.split('.')[0]
if drop_datatype:
txt = txt.replace('-parsed', '')
txt = txt.replace(' ', '_')
if not hyphens_ok:
txt = txt.replace('-', '_')
variable_safe = re.sub(variable_safe_r, '', txt)
if is_number(variable_safe):
variable_safe = 'c' + variable_safe
return variable_safe
def __getitem__(self, key):
"""allow slicing, indexing"""
from corpkit.process import makesafe
# allow slicing
if isinstance(key, slice):
n = OrderedDict()
for ii in range(*key.indices(len(self))):
n[self.keys()[ii]] = self[ii]
return Interrodict(n)
# allow integer index
elif isinstance(key, int):
return next(v for i, (k, v) in enumerate(self.items()) if i == key)
#return self.subcorpora.__getitem__(makesafe(self.subcorpora[key]))
# dict key access
else:
try:
return OrderedDict.__getitem__(self, key)
except:
from corpkit.process import is_number
if is_number(key):
return self.__getattribute__('c' + key)
def __getitem__(self, key):
"""allow slicing, indexing"""
from corpkit.process import makesafe
# allow slicing
if isinstance(key, slice):
n = OrderedDict()
for ii in range(*key.indices(len(self))):
n[self.keys()[ii]] = self[ii]
return Interrodict(n)
# allow integer index
elif isinstance(key, int):
return next(v for i, (k, v) in enumerate(self.items()) if i == key)
#return self.subcorpora.__getitem__(makesafe(self.subcorpora[key]))
# dict key access
else:
try:
return OrderedDict.__getitem__(self, key)
except:
from corpkit.process import is_number
if is_number(key):
return self.__getattribute__('c' + key)