* changed queue to set in AC3

Changed queue to set in AC3 (as in the pseudocode of the original algorithm) to reduce the number of consistency-check due to the redundancy of the same arcs in queue. For example, on the harder1 configuration of the Sudoku CSP the number consistency-check has been reduced from 40464 to 12562!

* re-added test commented by mistake

* Reworked PriorityQueue spec

Modified:
- Priority Queue methods:
    queue[elem] now returns the first value of elem stored in queue
    elem in queue now correctly returns whether a copy of element is present regardless of the function value. Apparently the bug was introduced while trying to meet heapq spec
    del queue[elem] deletes the first instance of elem in queue correctly
- Algorithms
    Same change in best_first_graph_search in romania_problem.py and search.py to make them compatible with the new spec
- Tests
    Introduced 3 tests in test_utils.py to comprehensively test PriorityQueue's new spec

* Reworked PriorityQueue spec

Modified:
- Priority Queue methods:
    queue[elem] now returns the first value of elem stored in queue
    elem in queue now correctly returns whether a copy of element is present regardless of the function value. Apparently the bug was introduced while trying to meet heapq spec
    del queue[elem] deletes the first instance of elem in queue correctly
- Algorithms
    Same change in best_first_graph_search in romania_problem.py and search.py to make them compatible with the new spec
- Tests
    Introduced 3 tests in test_utils.py to comprehensively test PriorityQueue's new spec

* update in Readme.md

* updated as per the suggestions in the review.

* Update in NeuralNetLearner function

* made the changes as suggested

* no need of if else

* no need of if - else .   

As if hidden_layer_sizes is zero then it will not affect layer_sizes

* removed comment

* Update learning.py

* update in multimap functi n in utils.py and added test for it

* made changes according to PEP style

* broke the long sentence to 2 shorter sentences

* updated the sequence function and added test cases

* updated according to the PEP style

* improved the count function in utils.py

* updated according to PEP style

* added some test cases to first function in utils.py

* added space after each comma

* Fix typos for Wupus agent

* fix imports

`usa` in csp.py incorrectly has NM not being a neighbor with AZ.

Otherwise `csp.curr_domains` may not be available for the loop that follows.

Shortened it, finding conflicts separately and storing them in different variables has no use later in the notebook; so i believe this looks better.