Skip to content
Branch: master
Find file Copy path
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
248 lines (178 sloc) 6.53 KB
# -*- coding: utf-8 -*-
import sys
import datetime
import re
import os
import logging
from decimal import Decimal, localcontext
from collections import namedtuple
from ordered_set import OrderedSet
logger = logging.getLogger(__name__)
py_major_version = sys.version[0]
py_minor_version = sys.version[2]
py2 = py_major_version == '2'
py3 = py_major_version == '3'
py4 = py_major_version == '4'
if py4:
logger.warning('Python 4 is not supported yet. Switching logic to Python 3.') # pragma: no cover
py3 = True # pragma: no cover
if py2: # pragma: no cover
sys.exit('Python 2 is not supported anymore. The last version of DeepDiff that supported Py2 was 3.3.0')
pypy3 = py3 and hasattr(sys, "pypy_translation_info")
# from builtins import int
strings = (str, bytes) # which are both basestring
unicode_type = str
bytes_type = bytes
numbers = (int, float, complex, datetime.datetime,, datetime.timedelta, Decimal)
IndexedHash = namedtuple('IndexedHash', 'indexes item')
current_dir = os.path.dirname(os.path.abspath(__file__))
ID_PREFIX = '!>*id'
KEY_TO_VAL_STR = "{}:{}"
def short_repr(item, max_length=15):
"""Short representation of item if it is too long"""
item = repr(item)
if len(item) > max_length:
item = '{}...{}'.format(item[:max_length - 3], item[-1])
return item
class ListItemRemovedOrAdded: # pragma: no cover
"""Class of conditions to be checked"""
class OtherTypes:
def __repr__(self):
return "Error: {}".format(self.__class__.__name__) # pragma: no cover
__str__ = __repr__
class Skipped(OtherTypes):
class Unprocessed(OtherTypes):
class NotHashed(OtherTypes):
class NotPresent: # pragma: no cover
In a change tree, this indicated that a previously existing object has been removed -- or will only be added
in the future.
We previously used None for this but this caused problem when users actually added and removed None. Srsly guys? :D
def __repr__(self):
return 'not present' # pragma: no cover
__str__ = __repr__
unprocessed = Unprocessed()
skipped = Skipped()
not_hashed = NotHashed()
notpresent = NotPresent()
# Disabling remapping from old to new keys since the mapping is deprecated.
RemapDict = dict
# class RemapDict(dict):
# """
# Remap Dictionary.
# For keys that have a new, longer name, remap the old key to the new key.
# Other keys that don't have a new name are handled as before.
# """
# def __getitem__(self, old_key):
# new_key = EXPANDED_KEY_MAP.get(old_key, old_key)
# if new_key != old_key:
# logger.warning(
# "DeepDiff Deprecation: %s is renamed to %s. Please start using "
# "the new unified naming convention.", old_key, new_key)
# if new_key in self:
# return self.get(new_key)
# else: # pragma: no cover
# raise KeyError(new_key)
class Verbose:
Global verbose level
level = 1
class indexed_set(set):
A set class that lets you get an item by index
>>> a = indexed_set()
>>> a.add(10)
>>> a.add(20)
>>> a[0]
Decimal: float,
OrderedSet: list,
type: lambda x: x.__name__,
bytes: lambda x: x.decode('utf-8')
def json_convertor_default(default_mapping=None):
_convertor_mapping = JSON_CONVERTOR.copy()
if default_mapping:
def _convertor(obj):
for original_type, convert_to in _convertor_mapping.items():
if isinstance(obj, original_type):
return convert_to(obj)
raise TypeError('We do not know how to convert {} of type {} for json serialization. Please pass the default_mapping parameter with proper mapping of the object to a basic python type.'.format(obj, type(obj)))
return _convertor
def add_to_frozen_set(parents_ids, item_id):
return parents_ids | {item_id}
def convert_item_or_items_into_set_else_none(items):
if items:
if isinstance(items, strings):
items = set([items])
items = set(items)
items = None
return items
RE_COMPILED_TYPE = type(re.compile(''))
def convert_item_or_items_into_compiled_regexes_else_none(items):
if items:
if isinstance(items, (strings, RE_COMPILED_TYPE)):
items = [items]
items = [i if isinstance(i, RE_COMPILED_TYPE) else re.compile(i) for i in items]
items = None
return items
def get_id(obj):
Adding some characters to id so they are not just integers to reduce the risk of collision.
return "{}{}".format(ID_PREFIX, id(obj))
def get_type(obj):
Get the type of object or if it is a class, return the class itself.
return obj if type(obj) is type else type(obj)
def type_in_type_group(item, type_group):
return get_type(item) in type_group
def type_is_subclass_of_type_group(item, type_group):
return isinstance(item, type_group) or issubclass(item, type_group) or type_in_type_group(item, type_group)
def get_doc(doc_filename):
with open(os.path.join(current_dir, doc_filename), 'r') as doc_file:
doc =
except Exception: # pragma: no cover
doc = 'Failed to load the docstrings. Please visit:' # pragma: no cover
return doc
number_formatting = {
"f": r'{:.%sf}',
"e": r'{:.%se}',
def number_to_string(number, significant_digits, number_format_notation="f"):
Convert numbers to string considering significant digits.
using = number_formatting[number_format_notation]
except KeyError:
raise ValueError("number_format_notation got invalid value of {}. The valid values are 'f' and 'e'".format(number_format_notation)) from None
if isinstance(number, Decimal):
tup = number.as_tuple()
with localcontext() as ctx:
ctx.prec = len(tup.digits) + tup.exponent + significant_digits
number = number.quantize(Decimal('0.' + '0' * significant_digits))
result = (using % significant_digits).format(number)
# Special case for 0: "-0.00" should compare equal to "0.00"
if set(result) <= ZERO_DECIMAL_CHARACTERS:
result = "0.00"
if number_format_notation == 'e' and isinstance(number, float):
result = result.replace('+0', '+')
return result
You can’t perform that action at this time.