This assignment focuses on implementing a Set ADT using binary search trees (BSTs) and AVL trees. Turn in
all parts of this assignment electronically. Your turnin should consist of four files: TrackingStreeSet.java,
AVLStreeSet.java, StringSetClient.java, and README.txt (this will contain your answers to the questions
found in Step 0 and 3 as described below). You will also need the additional supporting files: StringSet.java,
StreeNode.java, StreeSet.java, StringSetClient.java, AuthorChallenge.java, the two word lists, and the six
book files. All of these files can be found on the course website.
In lecture, we created a Set for Strings backed by a BST called StringTreeSet. In this assignment, we have
modified the name of this file and class to be StreeSet (String Tree Set) for brevity, but it is the same code we
discussed and developed together in lecture. In this assignment you will extend the functionality of our String
Set to keep track of the height of each node and maintain the AVL balance property for insertions and deletions.
Step 0: Warm-up
For the coding portions of this assignment, we encourage you to draw lots of examples of AVL trees to see how they work
for insertions and deletions. For this step, draw the AVL tree constructed by successive inserts of the following values:
2 1 4 5 9 3 6 7
You will not turn in your drawing, but give your answer to this step by saving the answers to the following questions in
1. Give the preorder traversal of your final tree.
2. Which values when inserted caused a single (outside) rotation?
3. Which values when inserted caused a double (inside) rotation?
Step 1: TrackingStreeSet to maintain height information
In this part of the assignment you will create a new class called TrackingStreeSet in a file named TrackingStreeSet.java. This class is a String Set backed by a BST that maintains height information for each node.
TrackingStreeSet should extend (i.e. be a subclass of) the provided StreeSet.
The provided StreeSet class has the following constructors and methods (along with a few others):
public StreeSet() constructs an empty tree set
public boolean add(String value) adds a value to the String set
public boolean contains(String value) returns true if the set contains the specified String
public boolean remove(String value) removes the specified String from this set if it is present
public int size() returns the number of elements in the set
public String toString() returns a String representation of StreeSet with elements in
their “natural order”
The version of StreeNode that you are receiving for this assignment has an additional field: height. This field
should be correctly maintained for each StreeNode in the TrackingStreeSet. Recall that the height of a tree is:
• -1 if it is empty,
• 0 if it has only one node, and
• the height of its tallest child plus 1 if there is more than one node.
One reason to store the height of a tree node is to compute a tree node’s balance factor. The balance factor for a
tree node, n, is the height of n’s right subtree minus the height of n’s left subtree.
2 of 5
Implement the following methods in TrackingStreeSet:
• protected static int computeHeight(StreeNode node)
This method returns the height of the given BST according to the definition of height outlined above.
This method should be implemented to run in constant time (i.e. use the nodes’ height fields rather
than actually traversing the tree).
• protected StreeNode add(StreeNode node, String value)
This method overrides the recursive helper add method found in the provided StreeSet. In this
method, you should add a value to the BST just as in the regular StreeSet and additionally maintain
the heights of nodes in the tree. Do not rewrite the add code found in the StreeSet class. TrackingStreeSet’s add should work exactly like StreeSet’s add except that after a node is added there is additional functionality to maintain heights of nodes. Therefore, your changes should only affect the
runtime of add by a constant (i.e. the runtime of this method should be O(log n)). The Java keyword
super is helpful here.
• protected StreeNode remove(StreeNode node, String value)
This method overrides the recursive helper remove method found in the provided StreeSet. In this
method, you should remove a value from the BST just as in the regular StreeSet and additionally
maintain the heights of nodes in the tree. Do not rewrite the remove code found in the StreeSet
class. TrackingStreeSet’s remove should work exactly like StreeSet’s remove except that after a
node is added there is additional functionality to maintain heights of nodes. Therefore, your changes
should only affect the runtime of remove by a constant (i.e. the runtime of this method should be
O(log n)). The Java keyword super is helpful here.
• protected String toString(StreeNode node)
This method overrides the recursive helper toString method found in the provided StreeSet. In this
method you should return a String representation of the BST that displays height information. For
example, if “a”, “b”, and “c” were inserted into a TrackingStreeSet in that order, the toString
method should return the String “[(a – 2), (b – 1), (c – 0)]”. The String returned should
be an in-order representation of the BST—start with a [, end with a ], and have a comma and a space
separating each node. A portion of the String that represents a node should be composed of an open
parenthesis, the value in the node, a space, a dash, another space, the height of the subtree rooted at the
node, and a closing parenthesis.
• protected static int balanceFactor(StreeNode node)
This method returns the balance factor of the given node. This method should be implemented to run
in constant time (i.e. use the nodes’ height fields rather than actually traversing the tree).
• public boolean isBalanced()
This method returns true if the following two conditions hold true for every node in the BST: (1) the
height of the node is correct, and (2) the balance factor of each node is either -1, 0, or 1. In other
words, this method returns true if the TrackingStreeSet complies with the AVL balance property
and false otherwise. Your implementation should run in O(n) time and not traverse any unnecessary
parts of the tree (i.e. once you know a section of the tree is unbalanced, there is no reason to check
branches that have not yet been traversed).
3 of 5
Step 2: AVLStreeSet to maintain balance property
In this part of the assignment you will create a new class called AVLStreeSet in a file named AVLStreeSet.java.
This class is a String Set backed by an AVL Tree that maintains the AVL balance property. AVLStreeSet
should extend (i.e. be a subclass of) your TrackingStreeSet.
An AVL Tree is a BST with an additional property: it maintains a balance factor of 0, 1, or -1 for each node.
When insertions/removals are done on an AVL tree, the AVL property could be violated. For a node, n, the
AVL property may be violated due to one of six cases:
1. LL Case: An insertion into / removal from the left subtree of the left child of n.
2. LR Case: An insertion into / removal from the right subtree of the left child of n.
3. RL Case: An insertion into / removal from the left subtree of the right child of n.
4. RR Case: An insertion into / removal from the right subtree of the right child of n.
5. -20 Case: A removal from the right subtree has occurred causing n to have a balance factor of -2 and the
left child of n has a balance factor of 0.
6. 20 Case: A removal from the left subtree has occurred causing n to have a balance factor of 2 and the
right child of n has a balance factor of 0.
Implement the following methods in AVLStreeSet:
• private StreeNode rightRotate(StreeNode parent)
This method should right rotate in order to fix the LL Case. (Hint: This code was given and explained
in lecture. Just add it to your AVLStreeSet file.)
• private StreeNode leftRotate(StreeNode parent)
This method should left rotate the parent, the parent’s right child, and the parent’s right child’s left subtree in order to fix the RR Case.
• private StreeNode rebalance(StreeNode node)
If the AVL Tree rooted at the given node is in violation of the AVL Tree balance condition, this
method should restore the AVL tree property for given node through rotations. (Hint: An outline of
this method was given and explained in lecture. Start with what we gave in lecture, modify it and add
code to take care of all of the different cases by which a node can become imbalanced.)
• protected StreeNode add(StreeNode node, String value)
This method overrides the recursive helper add method found in your TrackingStreeSet. In this
method you should add a value to the AVL Tree rooted at the given node just as in the TrackingStreeSet and rebalance the given node if the balance property is violated as a result of the addition.
Do not rewrite the add code found in your TrackingStreeSet. AVLStreeSet’s add should work like
TrackingStreeSet’s add except that after a value is added, the node may be rebalanced to maintain
the AVL Tree balance property. Your changes should only affect the runtime of add by a constant
(i.e. the runtime of this method should be O(log n)). The Java keyword super is helpful here.
• protected StreeNode remove(StreeNode node, String value)
This method overrides the recursive helper remove method found in your TrackingStreeSet. In this
method you should remove the value from the AVL Tree rooted at the given node just as in the
TrackingStreeSet and rebalance the given node if the balance property is violated as a result of the
removal. Do not rewrite the add code found in your TrackingStreeSet. AVLStreeSet’s remove
should work like TrackingStreeSet’s remove except that after a value is removed, the node may be
rebalanced to maintain the AVL Tree balance property. Your changes should only affect the runtime
of add by a constant (i.e. the runtime of this method should be O(log n)). The Java keyword super is
4 of 5
Step 3: Author Challenge!
There are many literary works that are considered to be of great importance. Amongst these are Cervante’s Don
Quixote (Spain), Shakespeare’s Hamlet (England), Joyce’s Ulysses (Ireland), and Tolstoy’s War and Peace
(Russia). In this part of the assignment, you will use our StreeSet and your AVLStreeSet to determine which
of these great works has the most diverse vocabulary using the provided AuthorChallenge.java code provided.
AuthorChallenge prompts the user if they would like to compare the books against a small list of words or a
large list of words and also for a book file (donquixote.txt, hamlet.txt, ulysses.txt, and warandpeace.txt have all been provided off of the course homework webpage). AuthorChallenge then builds two
sets: one using the provided StreeSet and another using your AVLStreeSet. AuthorChallenge determines the
percentage of the words used by the book by searching for every word in the chosen word list in both sets.
AuthorChallenge reports the time it takes to build each set and the time it took to search for all of the words.
Here’s an example log of execution:
Use (S)mall or (L)arge dictionary? l
Enter book file: ulysses.txt
BST Set build time: 190ms
AVL Set build time: 341ms
BST Set search time: 58ms (19.16% of the words are used)
AVL Set search time: 45ms (19.16% of the words are used)
Using AuthorChallenge, the provided StreeSet, and your AVLStreeSet, answer the following questions and
save them in your README.txt.
1. Run author challenge using the large word list and each of the four book files. Save each log of execution to your README.txt as the answer to this question. You might want to run AuthorChallenge several times for each set of inputs for a representative log of execution of running times.
2. Which author uses the most diverse vocabulary? In other words, which author has the highest percentage of words used?
3. Explain why your isBalanced method runs in O(n) time.
4. Which Set implementation takes longer to build? Why do you think this is the case?
5. Which Set implementation takes longer to search? Why do you think this is the case?
Development and Testing Strategy
We are providing you with a StringSetClient.java file that contains one method that tests an AVL Tree removal. We will not be providing you with any other test cases. You should add testing methods like the one
provided to this file and turn it in. You will be graded on how well you have tested your code. For example,
for TrackingStreeSet you should test for trees that are both balanced and imbalanced, and for your AVLStreeSet
you should test the all the different cases for AVL Tree inserts and removals.
We suggest that you do this assignment in the same order the steps and methods have been provided. After
writing each method, we suggest testing that method before moving on to writing other methods. It is easier to
find bugs after only a small amount of code has been written rather than waiting until you have entire classes
written. In the case of the rebalance method, you may want to write a single rebalance case and test that single case before moving onto writing the other rebalance cases.
Before you run AuthorChallenge on the large book files, you might want to try to run it on ulysses_frag.txt
and warandpeace_frag.txt to ensure that you are getting the right percentage of words used. It is almost impossible to know if your sets are working if you use the large word file and full book files right away.
5 of 5
Style Guidelines and Grading
A large part of your grade will come from appropriately utilizing a tree data structure to implement your
StreeSet.java and AVLStreeSet.java classes. There will be significant deductions if you use Java collections to implement these classes. Redundancy is another major grading focus; some methods are similar in behavior or based off of each other’s behavior. You should avoid repeated logic as much as possible. Your class
may have other methods besides those specified, but any other methods you add should be private.
You should follow good general style guidelines such as: making fields private and avoiding unnecessary
fields; appropriately using control structures like loops and if/else; properly using indentation, good variable
names and types; and not having any lines of code longer than 100 characters.
Comment your code descriptively in your own words at the top of your class, each method, and on complex
sections of your code. Comments should explain each method’s behavior, parameters, return, and exceptions.
The files provided to you use “doc comments” format used by Javadoc, but you do not have to do this. For reference, our solution for TrackingStreeSet.java is 51 lines long (32 lines if you ignore blank, closing braces,
and commented lines) and our solution for AVLStreeSet.java is 95 lines long (53 lines if you ignore blank,
closing braces, and commented lines).