Priority dict: a priority queue with updatable priorities You may have needed a priority queue with the ability to change the priorities of elements that were in the queue. This recipe solves this problem in an efficient way. The priority queue is implemented as a dictionary, where keys are the items of the queue, and values are their priorities…

priority_dict.py

# This is free and unencumbered software released into the public domain.
#
# Anyone is free to copy, modify, publish, use, compile, sell, or
# distribute this software, either in source code form or as a compiled
# binary, for any purpose, commercial or non-commercial, and by any
# means.
#
# In jurisdictions that recognize copyright laws, the author or authors
# of this software dedicate any and all copyright interest in the
# software to the public domain. We make this dedication for the benefit
# of the public at large and to the detriment of our heirs and
# successors. We intend this dedication to be an overt act of
# relinquishment in perpetuity of all present and future rights to this
# software under copyright law.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
# IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR
# OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
# ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
# OTHER DEALINGS IN THE SOFTWARE.
#
# For more information, please refer to <http://unlicense.org/>

# By Matteo Dell'Amico

from heapq import heapify, heappush, heappop

class priority_dict(dict):
    """Dictionary that can be used as a priority queue.

    Keys of the dictionary are items to be put into the queue, and values
    are their respective priorities. All dictionary methods work as expected.
    The advantage over a standard heapq-based priority queue is
    that priorities of items can be efficiently updated (amortized O(1))
    using code as 'thedict[item] = new_priority.'

    The 'smallest' method can be used to return the object with lowest
    priority, and 'pop_smallest' also removes it.

    The 'sorted_iter' method provides a destructive sorted iterator.
    """
    
    def __init__(self, *args, **kwargs):
        super(priority_dict, self).__init__(*args, **kwargs)
        self._rebuild_heap()

    def _rebuild_heap(self):
        self._heap = [(v, k) for k, v in self.iteritems()]
        heapify(self._heap)

    def smallest(self):
        """Return the item with the lowest priority.

        Raises IndexError if the object is empty.
        """
        
        heap = self._heap
        v, k = heap[0]
        while k not in self or self[k] != v:
            heappop(heap)
            v, k = heap[0]
        return k

    def pop_smallest(self):
        """Return the item with the lowest priority and remove it.

        Raises IndexError if the object is empty.
        """
        
        heap = self._heap
        v, k = heappop(heap)
        while k not in self or self[k] != v:
            v, k = heappop(heap)
        del self[k]
        return k

    def __setitem__(self, key, val):
        # We are not going to remove the previous value from the heap,
        # since this would have a cost O(n).
        
        super(priority_dict, self).__setitem__(key, val)
        
        if len(self._heap) < 2 * len(self):
            heappush(self._heap, (val, key))
        else:
            # When the heap grows larger than 2 * len(self), we rebuild it
            # from scratch to avoid wasting too much memory.
            self._rebuild_heap()

    def setdefault(self, key, val):
        if key not in self:
            self[key] = val
            return val
        return self[key]

    def update(self, *args, **kwargs):
        # Reimplementing dict.update is tricky -- see e.g.
        # http://mail.python.org/pipermail/python-ideas/2007-May/000744.html
        # We just rebuild the heap from scratch after passing to super.
        
        super(priority_dict, self).update(*args, **kwargs)
        self._rebuild_heap()

    def sorted_iter(self):
        """Sorted iterator of the priority dictionary items.

        Beware: this will destroy elements as they are returned.
        """
        
        while self:
            yield self.pop_smallest()