Project 4, Fake Min-Max Heaps

Due: Tuesday, April 25, 8:59pm

Addenda

[Apr 18 11:22] Fixed bug in sanity check should be Files have been updated. Please re-download test1.cpp, ..., test6.cpp.
[Apr 10 21:50] Additional implementation notes.
[Apr 10 21:45] Test files for checking copy constructor and assignment operator posted.

Corrections/clarifications in orange below.

Objectives

The objective of this programming assignment is to familiarize you with the binary heap data structure and for you to gain some experience designing templated C++ code.

Background

The premise of this project is that you've been asked, say by your boss, to implement a min-max heap. This data structure allows you to insert, delete the smallest item and delete the largest item all in O( log n ) time. From what you have heard about min-max heaps (e.g., Project 3 from the Fall 2013 CMSC 341), you know that they are difficult to implement and are chock-full of weird cases (q.v., Case 4). Suffice it to say, you really don't want to implement real min-max heaps.

So, you concoct a plan to implement a fake min-max heap. Why not just store every item in two heaps? a min-heap an a max-heap. Presumably insertion would still be in O( log n ) time, since you just insert twice. You could delete the smallest item from the min-heap and the largest item from the max-heap in O( log n ) time. You just have to store everything twice, but memory is cheap, right? Oh wait, when you delete the smallest item from the min-heap, its "twin" in the max-heap must also be removed. Since heaps do not support search, you'd have to link the "twins" somehow so when one is removed from one of the heaps, the other one can also be removed quickly.

Since your boss is only semi-literate in C++, you figure you can probably get away with this ruse --- you just can't be too obvious about it. For example, you can't have two separate classes, one for a min-heap and one for a max-heap. Also, having if statements that check a flag and then run the code for a min-heap versus a max-heap would be way too obvious. Since your boss wants you to make the code templated anyway, your plan is to make a templated heap class that will be instantiated once for a min-heap and another time for a max-heap.

Assignment

Note: Running time is one of the most important considerations in the implementation of a data structure. Programs that produce the desired output but exceed the required running times are considered wrong implementations and will receive substantial deductions during grading.

Your assignment is to design and implement a templated C++ class MinMaxHeap that supports insert, deleteMin and deleteMax in O( log n ) time. You are given general strategies and some specific requirements, but most of the class design and implementation is up to you.

The general requirements are:

MinMaxHeap must support insert, deleteMin and deleteMax in O( log n ) time. The restriction here is that searching for a key in a heap takes linear time. So, you can't delete the root of one heap and look for its twin in the other heap. That will take too long.
You may not use STL container classes. (Strings are OK.)
Your templated min-max heap must work with any class that supports assignment and comparison. (See test programs.)
You must have one implementation of a heap class that can be used as a min-heap and a max-heap. The code must use a function pointer to compare items. The behavior of this comparison function is what determines whether the heap is a min-heap or a max-heap.

General Strategies

You will need a wrapper class MinMaxHeap that has the functions used by the main program and a separate Heap class that does the "bubbling up" and "trickling down" to maintain a heap. A MinMaxHeap object will have two Heap objects --- one for the min-heap and one for the max-heap.
Each heap has to "know" that it is part of a min-max heap. It just doesn't "know" if it is the min-heap or the max-heap.
Each item in either the min-heap or the max-heap has to keep track of its "twin" in the other heap. The easiest way to do this is to store the index of the twin in the other heap. For example, consider the following output from the dump() function of a small min-max heap. The line means that key value 6 is stored in index 2 of the min-heap. The second value 5 indicates that key 6 is in index 5 of the max-heap. This confirmed by the output: in the max-heap. That is, the 5th item does indeed hold the key value 6. It also knows that in the min-heap, key value 6 is stored in index 2. You can confirm for each item in each heap that the index of the item's twin in the other heap is recorded properly. You can also confirm that the min-heap is heap ordered correctly with the smallest item in the root and the max-heap is heap ordered correctly with the largest itme int the root.
Every time an item is moved in one of the heaps, say in the bubbling up process after insertion, you have to update the item's twin so the twin "knows" the item's new position. (I.e., after moving an item, a dump should still be consistent as described above.) Look at the rest of the output from first test program for examples: test1.txt.
Each heap needs access to the internals of the other heap. So, in the MinMaxHeap constructor you have to create two heaps and then give each Heap a pointer to the other Heap.
You'll have to make templated stand-alone functions (i.e., not member functions) that compare two heap items and says which one is bigger or which one is smaller. You can assume that the type/class used to instantiate the templates supports the six comparison operators: <, <=, ==, !=, >= and >. The MinMaxHeap constructor should give a pointer to one of these comparison functions to the appropriate Heap object. That is what determines if the Heap is a max heap or a min heap.
The MinMaxHeap class has to coordinate insertion and deletion between the two heaps. For example, suppose that during insertion, you insert in the min-heap first and then insert in the max-heap. When you insert in the min-heap, it does not yet have a twin in the max-heap. After inserting in the min-heap, the MinMaxHeap insertion function has to know the location of the item in the array for the min-heap. This index has to be recorded with the item's twin in the max-heap. Then, the opposite has to happen: the newly inserted item's location in the max-heap must be recorded with its twin in the min-heap.
A similar coordination must also take place for deletion.
Your heaps have to support deleting an item given its index. (I.e., you have to be able to delete nodes other than the root.) The removed item should be replaced by the last item in the heap, as we've seen when the root is removed. However, when we fix the heap, we may have to bubble up or trickle down. (Re-read the last sentence.)

Specific Requirements

Your class definitions must be in a single file called MinMaxHeap.h (case sensitive). Client programs that use the MinMaxHeap class should be able to do so simply using: All of the template implementations should be in a file called MinMaxHeap.cpp (case sensitive).
The MinMaxHeap class must have a constructor with the signature:
The constructor must create a MinMaxHeap object capable of holding capacity items. The object created by the constructor must be able to accept insertion and deletion calls right after the constructor is done.
The MinMaxHeap class must have a copy constructor with the signature:
The constructor must create a copy of the MinMaxHeap object given in the parameter. The copied object must have its own allocated memory.
The MinMaxHeap class must have a destructor with the signature:
The destructor must deallocate any dynamically allocated memory.
The MinMaxHeap class must have an overloaded assignment operator with the signature:
The assignment operator must deallocate memory in the host object and copy the rhs MinMaxHeap into the host.
Your compiled code must run without memory leaks.
The MinMaxHeap class must have a member function called size that returns the number of items in the min-max heap. The size() function must have the signature:

The size() function must run constant time to receive full credit.
The MinMaxHeap class must have a member function called insert that adds the item given in the parameter to the min-max heap. The insert() function must have the signature:
If min-max heap cannot hold anymore items (i.e., it has reached capacity), then insert() should throw a out_of_range exception.
The insert() function must run in O(log n ) time to receive full credit.
The MinMaxHeap class must have a member function called deleteMin that removes the smallest item in the min-max heap and returns the key value of the deleted item. The deleteMin() function must have the signature:
If the heap is empty, the deleteMin() function should throw a out_of_range exception.
The deleteMin() function must run in O(log n ) time to receive full credit.
The MinMaxHeap class must have a member function called deleteMax that removes the largest item in the min-max heap and returns the key value of the deleted item. The deleteMax() function must have the signature:
If the heap is empty, the deleteMax() function should throw a out_of_range exception.
The deleteMax() function must run in O( log n ) time to receive full credit.
The MinMaxHeap class must have a member function called dump that prints out the contents of the min-max heap including the positions of each key in the min-heap and the max-heap along with recorded position of its "twin". See formatting examples in the sample output below. The dump() function must have the signature:
The MinMaxHeap class must have two member functions used for grading called locateMin and locateMax. These functions should have the prototype:
The function locateMin has two reference parameters for "return values". The data parameter should hold the value of the item in position pos of the min-heap. The index parameter should hold the location of the item's "twin" in the other heap.
The locateMax function have the analogous requirements.

Your code must run without segmentation fault and without memory leaks. For grading purposes, memory leaks are considered as bad as segmentation faults. This is because many segmentation faults are cause by poorly written destructors. A program with an empty destructor might avoid some segmentation faults but will leak memory horribly. Thus, not implementing a destructor or not deleting unused memory must incur a penalty that is equivalent to a segmentation fault.

Test Programs

Here are sample driver programs to test your implementation. Passing these tests do not mean you will receive 100% on your project. It does not guarantee that you will pass tests used in grading. You should make additional tests of your own!

Test MinMaxHeap with int data. A sanityCheck() function is provided. Compile using: Driver program: test1.cpp
Sample Output: test1.txt
Test MinMaxHeap with the string data. A sanityCheck() function is provided. Compile using: Driver program: test2.cpp
Sample output: test2.txt
Test MinMaxHeap with the NoCopyString class. A sanityCheck() function is provided. Compile using: Driver program: test3.cpp,
NoCopyString class: NoCopyString.h, NoCopyString.cpp,
Sample output: test3.txt
Bigger test of MinMaxHeap with int data. A sanityCheck() function is provided. The keys in the MinMaxHeap are also checked against an equivalent STL multiset. Only deleteMin() is checked because STL multiset::erase() doesn't want to erase reverse iterators. Compile using: Driver program: test4.cpp
Sample output: test4.txt
Same as previous test, except it checks deleteMax() instead of deleteMin(). Compile using: Driver program: test5.cpp
Sample output: test5.txt
Test copy constructor, destructor and assignment operator Should test this with valgrind
Driver program: test6.cpp
Sample output: test6.txt

Implementation Notes

Here we list some recommendations and point out some traps and pitfalls.

Review C++ template syntax before starting. Remember that you should
at the bottom of the header file. Also, the .cpp file must also be guarded with #ifndef.
Just start at index 1 for the heaps. The "wasted" storage in the 0-th index could be useful (hint, hint).
Beware of off-by-one errors since the 0-th item of the heap array is not used. A heap with capacity for 100 items needs 101 entries indexed 0 thru 100.
After you delete an item from a heap, the "twin" object has an invalid index for its twin.
There is a special case for deleting the last item of a heap. This is because we would usually swap in the last item to replace the item that is deleted. However, if are deleting the last item, then we would be swapping into an invalid location.
When a MinMaxHeap is copied, you must make sure that in the new MinMaxHeap, the new min-heap and the new max-heap are referencing each other and not the max-heap and min-heap of the original MinMaxHeap. The same situation applies for assignment.

What to Submit

You must submit the following files to the proj4 directory.

MinMaxHeap.h
MinMaxHeap.cpp
Driver.cpp

The Driver.cpp program should include tests showing the parts of your project that work correctly.

If you followed the instructions in the Project Submission page to set up your directories, you can submit your code using this Unix command command.

cp MinMaxHeap.h MinMaxHeap.cpp Driver.cpp ~/cs341proj/proj4/

CMSC 341 Data Structures — All Sections — Spring 2017