# Copyright (c) 2013, Mahmoud Hashemi # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions are # met: # # * Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # # * Redistributions in binary form must reproduce the above # copyright notice, this list of conditions and the following # disclaimer in the documentation and/or other materials provided # with the distribution. # # * The names of the contributors may not be used to endorse or # promote products derived from this software without specific # prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS # "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT # LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR # A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT # OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, # SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT # LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, # DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY # THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. """\ The :class:`set` type brings the practical expressiveness of set theory to Python. It has a very rich API overall, but lacks a couple of fundamental features. For one, sets are not ordered. On top of this, sets are not indexable, i.e, ``my_set[8]`` will raise an :exc:`TypeError`. The :class:`IndexedSet` type remedies both of these issues without compromising on the excellent complexity characteristics of Python's built-in set implementation. """ from bisect import bisect_left from collections.abc import MutableSet from itertools import chain, islice import operator try: from .typeutils import make_sentinel _MISSING = make_sentinel(var_name='_MISSING') except ImportError: _MISSING = object() __all__ = ['IndexedSet', 'complement'] _COMPACTION_FACTOR = 8 # TODO: inherit from set() # TODO: .discard_many(), .remove_many() # TODO: raise exception on non-set params? # TODO: technically reverse operators should probably reverse the # order of the 'other' inputs and put self last (to try and maintain # insertion order) class IndexedSet(MutableSet): """``IndexedSet`` is a :class:`collections.MutableSet` that maintains insertion order and uniqueness of inserted elements. It's a hybrid type, mostly like an OrderedSet, but also :class:`list`-like, in that it supports indexing and slicing. Args: other (iterable): An optional iterable used to initialize the set. >>> x = IndexedSet(list(range(4)) + list(range(8))) >>> x IndexedSet([0, 1, 2, 3, 4, 5, 6, 7]) >>> x - set(range(2)) IndexedSet([2, 3, 4, 5, 6, 7]) >>> x[-1] 7 >>> fcr = IndexedSet('freecreditreport.com') >>> ''.join(fcr[:fcr.index('.')]) 'frecditpo' Standard set operators and interoperation with :class:`set` are all supported: >>> fcr & set('cash4gold.com') IndexedSet(['c', 'd', 'o', '.', 'm']) As you can see, the ``IndexedSet`` is almost like a ``UniqueList``, retaining only one copy of a given value, in the order it was first added. For the curious, the reason why IndexedSet does not support setting items based on index (i.e, ``__setitem__()``), consider the following dilemma:: my_indexed_set = [A, B, C, D] my_indexed_set[2] = A At this point, a set requires only one *A*, but a :class:`list` would overwrite *C*. Overwriting *C* would change the length of the list, meaning that ``my_indexed_set[2]`` would not be *A*, as expected with a list, but rather *D*. So, no ``__setitem__()``. Otherwise, the API strives to be as complete a union of the :class:`list` and :class:`set` APIs as possible. """ def __init__(self, other=None): self.item_index_map = dict() self.item_list = [] self.dead_indices = [] self._compactions = 0 self._c_max_size = 0 if other: self.update(other) # internal functions @property def _dead_index_count(self): return len(self.item_list) - len(self.item_index_map) def _compact(self): if not self.dead_indices: return self._compactions += 1 dead_index_count = self._dead_index_count items, index_map = self.item_list, self.item_index_map self._c_max_size = max(self._c_max_size, len(items)) for i, item in enumerate(self): items[i] = item index_map[item] = i del items[-dead_index_count:] del self.dead_indices[:] def _cull(self): ded = self.dead_indices if not ded: return items, ii_map = self.item_list, self.item_index_map if not ii_map: del items[:] del ded[:] elif len(ded) > 384: self._compact() elif self._dead_index_count > (len(items) / _COMPACTION_FACTOR): self._compact() elif items[-1] is _MISSING: # get rid of dead right hand side num_dead = 1 while items[-(num_dead + 1)] is _MISSING: num_dead += 1 if ded and ded[-1][1] == len(items): del ded[-1] del items[-num_dead:] def _get_real_index(self, index): if index < 0: index += len(self) if not self.dead_indices: return index real_index = index for d_start, d_stop in self.dead_indices: if real_index < d_start: break real_index += d_stop - d_start return real_index def _get_apparent_index(self, index): if index < 0: index += len(self) if not self.dead_indices: return index apparent_index = index for d_start, d_stop in self.dead_indices: if index < d_start: break apparent_index -= d_stop - d_start return apparent_index def _add_dead(self, start, stop=None): # TODO: does not handle when the new interval subsumes # multiple existing intervals dints = self.dead_indices if stop is None: stop = start + 1 cand_int = [start, stop] if not dints: dints.append(cand_int) return int_idx = bisect_left(dints, cand_int) dint = dints[int_idx - 1] d_start, d_stop = dint if start <= d_start <= stop: dint[0] = start elif start <= d_stop <= stop: dint[1] = stop else: dints.insert(int_idx, cand_int) return # common operations (shared by set and list) def __len__(self): return len(self.item_index_map) def __contains__(self, item): return item in self.item_index_map def __iter__(self): return (item for item in self.item_list if item is not _MISSING) def __reversed__(self): item_list = self.item_list return (item for item in reversed(item_list) if item is not _MISSING) def __repr__(self): return f'{self.__class__.__name__}({list(self)!r})' def __eq__(self, other): if isinstance(other, IndexedSet): return len(self) == len(other) and list(self) == list(other) try: return set(self) == set(other) except TypeError: return False @classmethod def from_iterable(cls, it): "from_iterable(it) -> create a set from an iterable" return cls(it) # set operations def add(self, item): "add(item) -> add item to the set" if item not in self.item_index_map: self.item_index_map[item] = len(self.item_list) self.item_list.append(item) def remove(self, item): "remove(item) -> remove item from the set, raises if not present" try: didx = self.item_index_map.pop(item) except KeyError: raise KeyError(item) self.item_list[didx] = _MISSING self._add_dead(didx) self._cull() def discard(self, item): "discard(item) -> discard item from the set (does not raise)" try: self.remove(item) except KeyError: pass def clear(self): "clear() -> empty the set" del self.item_list[:] del self.dead_indices[:] self.item_index_map.clear() def isdisjoint(self, other): "isdisjoint(other) -> return True if no overlap with other" iim = self.item_index_map for k in other: if k in iim: return False return True def issubset(self, other): "issubset(other) -> return True if other contains this set" if len(other) < len(self): return False for k in self.item_index_map: if k not in other: return False return True def issuperset(self, other): "issuperset(other) -> return True if set contains other" if len(other) > len(self): return False iim = self.item_index_map for k in other: if k not in iim: return False return True def union(self, *others): "union(*others) -> return a new set containing this set and others" return self.from_iterable(chain(self, *others)) def iter_intersection(self, *others): "iter_intersection(*others) -> iterate over elements also in others" for k in self: for other in others: if k not in other: break else: yield k return def intersection(self, *others): "intersection(*others) -> get a set with overlap of this and others" if len(others) == 1: other = others[0] return self.from_iterable(k for k in self if k in other) return self.from_iterable(self.iter_intersection(*others)) def iter_difference(self, *others): "iter_difference(*others) -> iterate over elements not in others" for k in self: for other in others: if k in other: break else: yield k return def difference(self, *others): "difference(*others) -> get a new set with elements not in others" if len(others) == 1: other = others[0] return self.from_iterable(k for k in self if k not in other) return self.from_iterable(self.iter_difference(*others)) def symmetric_difference(self, *others): "symmetric_difference(*others) -> XOR set of this and others" ret = self.union(*others) return ret.difference(self.intersection(*others)) __or__ = __ror__ = union __and__ = __rand__ = intersection __sub__ = difference __xor__ = __rxor__ = symmetric_difference def __rsub__(self, other): vals = [x for x in other if x not in self] return type(other)(vals) # in-place set operations def update(self, *others): "update(*others) -> add values from one or more iterables" if not others: return # raise? elif len(others) == 1: other = others[0] else: other = chain(others) for o in other: self.add(o) def intersection_update(self, *others): "intersection_update(*others) -> discard self.difference(*others)" for val in self.difference(*others): self.discard(val) def difference_update(self, *others): "difference_update(*others) -> discard self.intersection(*others)" if self in others: self.clear() for val in self.intersection(*others): self.discard(val) def symmetric_difference_update(self, other): # note singular 'other' "symmetric_difference_update(other) -> in-place XOR with other" if self is other: self.clear() for val in other: if val in self: self.discard(val) else: self.add(val) def __ior__(self, *others): self.update(*others) return self def __iand__(self, *others): self.intersection_update(*others) return self def __isub__(self, *others): self.difference_update(*others) return self def __ixor__(self, *others): self.symmetric_difference_update(*others) return self def iter_slice(self, start, stop, step=None): "iterate over a slice of the set" iterable = self if start is not None: start = self._get_real_index(start) if stop is not None: stop = self._get_real_index(stop) if step is not None and step < 0: step = -step iterable = reversed(self) return islice(iterable, start, stop, step) # list operations def __getitem__(self, index): try: start, stop, step = index.start, index.stop, index.step except AttributeError: index = operator.index(index) else: iter_slice = self.iter_slice(start, stop, step) return self.from_iterable(iter_slice) if index < 0: index += len(self) real_index = self._get_real_index(index) try: ret = self.item_list[real_index] except IndexError: raise IndexError('IndexedSet index out of range') return ret def pop(self, index=None): "pop(index) -> remove the item at a given index (-1 by default)" item_index_map = self.item_index_map len_self = len(item_index_map) if index is None or index == -1 or index == len_self - 1: ret = self.item_list.pop() del item_index_map[ret] else: real_index = self._get_real_index(index) ret = self.item_list[real_index] self.item_list[real_index] = _MISSING del item_index_map[ret] self._add_dead(real_index) self._cull() return ret def count(self, val): "count(val) -> count number of instances of value (0 or 1)" if val in self.item_index_map: return 1 return 0 def reverse(self): "reverse() -> reverse the contents of the set in-place" reversed_list = list(reversed(self)) self.item_list[:] = reversed_list for i, item in enumerate(self.item_list): self.item_index_map[item] = i del self.dead_indices[:] def sort(self, **kwargs): "sort() -> sort the contents of the set in-place" sorted_list = sorted(self, **kwargs) if sorted_list == self.item_list: return self.item_list[:] = sorted_list for i, item in enumerate(self.item_list): self.item_index_map[item] = i del self.dead_indices[:] def index(self, val): "index(val) -> get the index of a value, raises if not present" try: return self._get_apparent_index(self.item_index_map[val]) except KeyError: cn = self.__class__.__name__ raise ValueError(f'{val!r} is not in {cn}') def complement(wrapped): """Given a :class:`set`, convert it to a **complement set**. Whereas a :class:`set` keeps track of what it contains, a `complement set `_ keeps track of what it does *not* contain. For example, look what happens when we intersect a normal set with a complement set:: >>> list(set(range(5)) & complement(set([2, 3]))) [0, 1, 4] We get the everything in the left that wasn't in the right, because intersecting with a complement is the same as subtracting a normal set. Args: wrapped (set): A set or any other iterable which should be turned into a complement set. All set methods and operators are supported by complement sets, between other :func:`complement`-wrapped sets and/or regular :class:`set` objects. Because a complement set only tracks what elements are *not* in the set, functionality based on set contents is unavailable: :func:`len`, :func:`iter` (and for loops), and ``.pop()``. But a complement set can always be turned back into a regular set by complementing it again: >>> s = set(range(5)) >>> complement(complement(s)) == s True .. note:: An empty complement set corresponds to the concept of a `universal set `_ from mathematics. Complement sets by example ^^^^^^^^^^^^^^^^^^^^^^^^^^ Many uses of sets can be expressed more simply by using a complement. Rather than trying to work out in your head the proper way to invert an expression, you can just throw a complement on the set. Consider this example of a name filter:: >>> class NamesFilter(object): ... def __init__(self, allowed): ... self._allowed = allowed ... ... def filter(self, names): ... return [name for name in names if name in self._allowed] >>> NamesFilter(set(['alice', 'bob'])).filter(['alice', 'bob', 'carol']) ['alice', 'bob'] What if we want to just express "let all the names through"? We could try to enumerate all of the expected names:: ``NamesFilter({'alice', 'bob', 'carol'})`` But this is very brittle -- what if at some point over this object is changed to filter ``['alice', 'bob', 'carol', 'dan']``? Even worse, what about the poor programmer who next works on this piece of code? They cannot tell whether the purpose of the large allowed set was "allow everything", or if 'dan' was excluded for some subtle reason. A complement set lets the programmer intention be expressed succinctly and directly:: NamesFilter(complement(set())) Not only is this code short and robust, it is easy to understand the intention. """ if type(wrapped) is _ComplementSet: return wrapped.complemented() if type(wrapped) is frozenset: return _ComplementSet(excluded=wrapped) return _ComplementSet(excluded=set(wrapped)) def _norm_args_typeerror(other): '''normalize args and raise type-error if there is a problem''' if type(other) in (set, frozenset): inc, exc = other, None elif type(other) is _ComplementSet: inc, exc = other._included, other._excluded else: raise TypeError('argument must be another set or complement(set)') return inc, exc def _norm_args_notimplemented(other): '''normalize args and return NotImplemented (for overloaded operators)''' if type(other) in (set, frozenset): inc, exc = other, None elif type(other) is _ComplementSet: inc, exc = other._included, other._excluded else: return NotImplemented, None return inc, exc class _ComplementSet: """ helper class for complement() that implements the set methods """ __slots__ = ('_included', '_excluded') def __init__(self, included=None, excluded=None): if included is None: assert type(excluded) in (set, frozenset) elif excluded is None: assert type(included) in (set, frozenset) else: raise ValueError('one of included or excluded must be a set') self._included, self._excluded = included, excluded def __repr__(self): if self._included is None: return f'complement({repr(self._excluded)})' return f'complement(complement({repr(self._included)}))' def complemented(self): '''return a complement of the current set''' if type(self._included) is frozenset or type(self._excluded) is frozenset: return _ComplementSet(included=self._excluded, excluded=self._included) return _ComplementSet( included=None if self._excluded is None else set(self._excluded), excluded=None if self._included is None else set(self._included)) __invert__ = complemented def complement(self): '''convert the current set to its complement in-place''' self._included, self._excluded = self._excluded, self._included def __contains__(self, item): if self._included is None: return not item in self._excluded return item in self._included def add(self, item): if self._included is None: if item in self._excluded: self._excluded.remove(item) else: self._included.add(item) def remove(self, item): if self._included is None: self._excluded.add(item) else: self._included.remove(item) def pop(self): if self._included is None: raise NotImplementedError # self.missing.add(random.choice(gc.objects())) return self._included.pop() def intersection(self, other): try: return self & other except NotImplementedError: raise TypeError('argument must be another set or complement(set)') def __and__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + return _ComplementSet(included=inc - self._excluded) else: # - - return _ComplementSet(excluded=self._excluded.union(other._excluded)) else: if inc is None: # + - return _ComplementSet(included=exc - self._included) else: # + + return _ComplementSet(included=self._included.intersection(inc)) __rand__ = __and__ def __iand__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + self._excluded = inc - self._excluded # TODO: do this in place? else: # - - self._excluded |= exc else: if inc is None: # + - self._included -= exc self._included, self._excluded = None, self._included else: # + + self._included &= inc return self def union(self, other): try: return self | other except NotImplementedError: raise TypeError('argument must be another set or complement(set)') def __or__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + return _ComplementSet(excluded=self._excluded - inc) else: # - - return _ComplementSet(excluded=self._excluded.intersection(exc)) else: if inc is None: # + - return _ComplementSet(excluded=exc - self._included) else: # + + return _ComplementSet(included=self._included.union(inc)) __ror__ = __or__ def __ior__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + self._excluded -= inc else: # - - self._excluded &= exc else: if inc is None: # + - self._included, self._excluded = None, exc - self._included # TODO: do this in place? else: # + + self._included |= inc return self def update(self, items): if type(items) in (set, frozenset): inc, exc = items, None elif type(items) is _ComplementSet: inc, exc = items._included, items._excluded else: inc, exc = frozenset(items), None if self._included is None: if exc is None: # - + self._excluded &= inc else: # - - self._excluded.discard(exc) else: if inc is None: # + - self._included &= exc self._included, self._excluded = None, self._excluded else: # + + self._included.update(inc) def discard(self, items): if type(items) in (set, frozenset): inc, exc = items, None elif type(items) is _ComplementSet: inc, exc = items._included, items._excluded else: inc, exc = frozenset(items), None if self._included is None: if exc is None: # - + self._excluded.update(inc) else: # - - self._included, self._excluded = exc - self._excluded, None else: if inc is None: # + - self._included &= exc else: # + + self._included.discard(inc) def symmetric_difference(self, other): try: return self ^ other except NotImplementedError: raise TypeError('argument must be another set or complement(set)') def __xor__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + return _ComplementSet(excluded=self._excluded - inc) else: # - - return _ComplementSet(included=self._excluded.symmetric_difference(exc)) else: if inc is None: # + - return _ComplementSet(excluded=exc - self._included) else: # + + return _ComplementSet(included=self._included.symmetric_difference(inc)) __rxor__ = __xor__ def symmetric_difference_update(self, other): inc, exc = _norm_args_typeerror(other) if self._included is None: if exc is None: # - + self._excluded |= inc else: # - - self._excluded.symmetric_difference_update(exc) self._included, self._excluded = self._excluded, None else: if inc is None: # + - self._included |= exc self._included, self._excluded = None, self._included else: # + + self._included.symmetric_difference_update(inc) def isdisjoint(self, other): inc, exc = _norm_args_typeerror(other) if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + return inc.issubset(self._excluded) else: # - - return False else: if inc is None: # + - return self._included.issubset(exc) else: # + + return self._included.isdisjoint(inc) def issubset(self, other): '''everything missing from other is also missing from self''' try: return self <= other except NotImplementedError: raise TypeError('argument must be another set or complement(set)') def __le__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + return False else: # - - return self._excluded.issupserset(exc) else: if inc is None: # + - return self._included.isdisjoint(exc) else: # + + return self._included.issubset(inc) def __lt__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + return False else: # - - return self._excluded > exc else: if inc is None: # + - return self._included.isdisjoint(exc) else: # + + return self._included < inc def issuperset(self, other): '''everything missing from self is also missing from super''' try: return self >= other except NotImplementedError: raise TypeError('argument must be another set or complement(set)') def __ge__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + return not self._excluded.intersection(inc) else: # - - return self._excluded.issubset(exc) else: if inc is None: # + - return False else: # + + return self._included.issupserset(inc) def __gt__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + return not self._excluded.intersection(inc) else: # - - return self._excluded < exc else: if inc is None: # + - return False else: # + + return self._included > inc def difference(self, other): try: return self - other except NotImplementedError: raise TypeError('argument must be another set or complement(set)') def __sub__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + return _ComplementSet(excluded=self._excluded | inc) else: # - - return _ComplementSet(included=exc - self._excluded) else: if inc is None: # + - return _ComplementSet(included=self._included & exc) else: # + + return _ComplementSet(included=self._included.difference(inc)) def __rsub__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented # rsub, so the expression being evaluated is "other - self" if self._included is None: if exc is None: # - + return _ComplementSet(included=inc & self._excluded) else: # - - return _ComplementSet(included=self._excluded - exc) else: if inc is None: # + - return _ComplementSet(excluded=exc | self._included) else: # + + return _ComplementSet(included=inc.difference(self._included)) def difference_update(self, other): try: self -= other except NotImplementedError: raise TypeError('argument must be another set or complement(set)') def __isub__(self, other): inc, exc = _norm_args_notimplemented(other) if inc is NotImplemented: return NotImplemented if self._included is None: if exc is None: # - + self._excluded |= inc else: # - - self._included, self._excluded = exc - self._excluded, None else: if inc is None: # + - self._included &= exc else: # + + self._included.difference_update(inc) return self def __eq__(self, other): return ( type(self) is type(other) and self._included == other._included and self._excluded == other._excluded) or ( type(other) in (set, frozenset) and self._included == other) def __hash__(self): return hash(self._included) ^ hash(self._excluded) def __len__(self): if self._included is not None: return len(self._included) raise NotImplementedError('complemented sets have undefined length') def __iter__(self): if self._included is not None: return iter(self._included) raise NotImplementedError('complemented sets have undefined contents') def __bool__(self): if self._included is not None: return bool(self._included) return True