Project 2AB: KDTree and Extrinsic PQ

Over the next 5 weeks, you are going to building a system very similar to Google Maps.

This system will have four major parts:

Projects 2A and 2B are being released at the same time and in the same spec since they are quite similar in structure.

All parts of this project should be completed individually. Please refer to the course policies for our official policy on academic dishonesty.

This part of the project will be as much about testing as it is about implementation. The autograder will be very sparse, and it will be up to you to test your own code! Grading details for project 2A and 2B will not be released until the amnesty period for submissions is over (December 12th).

In this document, we’ll only discuss proj2ab. More details about the mapping system (proj2c and proj2d) will come later.

Proj2A: K-d Tree

For this part of the project, your primary goal will be to implement the KDTree class. This class implements a k-d tree as described in lecture 19. This part of the project is a warmup to part 2B, where you’ll be doing something quite similar, but with much less guidance from us.

If you’re not familiar with how k-d trees work, please see Lecture 19 before proceeding. It is absolutely imperative that you do this before starting this assignment.

This will ultimately be useful in your BearMaps application. Specifically, a user will click on a point on the map, and your k-d tree will be used to find the nearest intersection to that click.

As mentioned above, the autograder for this project is minimal. Thus, you’ll need to verify the correctness and speed of your code on your own using your own tests. Since this is probably the first time you’ve ever had to do this in a project, I’ve created a pseudo-walkthrough video where I give specific recommendations for steps that you might want to take to complete the project. This is ONLY available for project 2a! When you get to project 2b, you’ll be entirely on your own.

IMPORTANT NOTE: YOU MAY NOT SHARE ANY CODE ON THIS ASSIGNMENT WITH ANYONE. It is not OK to share code, even if it’s very similar to the code shown in the walkthrough video.

This walkthrough also includes links to a series of solutions videos, where in each video, I solve one part of the project (with the code mostly blurred out). These videos cover some, but not all of the project. Even if you solve the project entirely on your own, you still might find these solutions videos useful, so that you can compare your problem solving strategy to mine.

Slides for this pseudo-walkthrough (including links to solution videos) can be found at this link. You are not required to follow these steps or use this walkthrough!

Provided Files

We provide a class called that represents a location with an x and y coordinate. You may not modify this class. There are three useful methods in this class:

We also provide a interface called PointSet that represents a set of such points. It has only one method.


As we saw in lab 5, it is often easiest to test the implementation of a complex but fast data structure by first implementing a simple but slow data structure that we can use as a gold standard. Before you create your efficient KDTree class, you will first create a naive linear-time solution to solve the problem of finding the closest point to a given coordinate. The goal of creating this class is that you will have a alternative, albeit slower, solution that you can use to easily verify if the result of your k-d tree’s nearest is correct. Create a class called NaivePointSet The declaration on your class should be:

    public class NaivePointSet implements PointSet {

Your NaivePointSet should have the following constructor and method:

Note that your NaivePointSet class should be immutable, meaning that you cannot add or or remove points from it. You do not need to do anything special to guarantee this.

Example Usage

    Point p1 = new Point(1.1, 2.2); // constructs a Point with x = 1.1, y = 2.2
    Point p2 = new Point(3.3, 4.4);
    Point p3 = new Point(-2.9, 4.2);
    NaivePointSet nn = new NaivePointSet(List.of(p1, p2, p3));
    Point ret = nn.nearest(3.0, 4.0); // returns p2
    ret.getX(); // evaluates to 3.3
    ret.getY(); // evaluates to 4.4

Part 2: K-d Tree

Now, create the class KDTree class. Your KDTree should have the following constructor and method:

As with NaivePointSet, your KDTree class should be immutable. Also note that while k-d trees can theoretically work for any number of dimensions, your implementation only has to work for the 2-dimensional case, i.e. when our points have only x and y coordinates.

For nearest, we recommend that you write a simple version and verify that it works before adding in the various optimizations from the pseudocode. For example, you might start by writing a nearest method that simply traverses your entire tree. Then, you might add code so that it always visits the “good” child before the “bad” child. Then after verifying that this works, you might try writing code that prunes the bad side. By doing things this way, you’re testing smaller ideas at once.


There are a number of different ways that you can construct tests for this part of the project, but we will go over our recommended approach.

It’s imporant to note that the Points that your KDTree and NaivePointSet’s return do not need to be exactly the same: instead, you should check that the distance from the goal Point to each of the Points that your KDTree and NaivePointSet finds are equal. If you’re using JUnit (which you should be), then you should read the JUnit javadoc for the assertEquals method for double values. You’ll see that you actually need to pass in a third argument which you should think of as a threshold. Use a very small threshold like 0.00000001.

Basic Sanity Checks

After you are confident in your partial implementation (the constructor or unoptimized nearest, for example), in a main method, construct the k-d tree from lecture 19, and verify with the Java Visualizer that you are able to construct the tree correctly, and return the right Point for a particular nearest query. You are likely to gain some confidence in your implementation through this exercise, and also discover some bugs. When you are done with your sanity checks, you can move forward to more rigorous randomized testing.

Randomized Testing

One approach to testing this method would be to create a bunch of hand-curated datasets to pass to the constructor and calls to nearest that try to verify various edge cases. To avoid thinking about all possible strange edge cases, we can turn towards techniques other than creating specific, deterministic tests to cover all of the possible errors.

Our suggestion is to use the lab 5 approach of using randomized comparison tests which will allow you to test your code on a large sample of points which should encompass most if not all possible edge cases.

As in lab 5, we recommend generating a large number of random points to be in your tree, as well as a large number of points to query using the nearest function. To verify the correctness of the results you should be able to compare the results of your KDTree’s nearest function to the results to the NaivePointSet’s nearest function. If we test a large number of queries without error we can be fairly confident in the correctness of our data structure and algorithm.

An issue is that randomized tests are not deterministic. This mean if we run the tests again and again, different points will be generated each time which will make debugging nearly impossible because we do not have any ability to replicate the failure cases. However, randomness in computers is almost never true randomness and is instead generated by pseudorandom number generators (PRNGs). Tests can be made deterministic by seeding the PRNG, where we are essentially initializing the PRNG to start in a specific state such that the random numbers generated will be the same each time. We suggest you use the class Random which will allow you to generate random coordinate values as well as provide the seed for the PRNG. More can be found about the specifications of this class in the online documentation.

Please put any tests you write in You cannot share your tests with other students, but you are free to discuss testing strategies with them. Even if your tests are very similar to the ones I write in my video, YOU MAY STILL NOT SHARE YOUR CODE!

Note: Our autograder will test points that have x and y coordinates whose magnitudes are less than 1000. We recommend your tests also consider points over this range.

Timing Tests

You will also be required to verify the speed of your k-d tree.

It should be possible to build a k-d tree of a million random points in a reasonable amount of time (less than 20 seconds). For example, our KDTree consruction timing table is displayed below. Note, this was running on a fairly fast computer, so your times may be longer. If you can hit 1,000,000 points in less than 20 seconds, you’re in good shape. Don’t worry about worst case inputs (i.e. your code should not gracefully handle a very spindly k-d tree, this requires a much more sophisticated approach, e.g. a k-d-b-tree). Note that “# ops” here refers to the number of points added to the table, and doesn’t actually represent the count of any specific method calls.

Timing table for Kd-Tree Construction
           N     time (s)        # ops  microsec/op
       31250         0.03        31250         1.02
       62500         0.05        62500         0.78
      125000         0.10       125000         0.82
      250000         0.23       250000         0.90
      500000         0.63       500000         1.26
     1000000         1.25      1000000         1.25
     2000000         3.14      2000000         1.57

Your KDTree’s nearest method should also be MUCH faster than your NaivePointSet. For example, our timing table for 1,000,000 queries on a NaivePointSet with N points is given below:

Timing table for Naive Nearest
           N     time (s)        # ops  microsec/op
         125         0.30      1000000         0.30
         250         0.47      1000000         0.47
         500         0.80      1000000         0.80
        1000         1.66      1000000         1.66

And for our KDTree below. Note that our KDTree was able to handle PointSets that were 1,000 times as large as a a naive implementation.

Timing table for Kd-Tree Nearest
           N     time (s)        # ops  microsec/op
       31250         0.93      1000000         0.93
       62500         0.86      1000000         0.86
      125000         1.03      1000000         1.03
      250000         1.34      1000000         1.34
      500000         2.09      1000000         2.09
     1000000         2.38      1000000         2.38

Note: To use the Stopwatch class you’ll need to import edu.princeton.cs.algs4.Stopwatch;.

External Resources

If you don’t like our presentation of the k-d tree, you are free to reference pseudocode from course slides and other websites (use the @source tag to annotate your sources), but all code you write must be your own. We strongly recommend that you use the approach described in CS 61B, as it is simpler than described elsewhere.

You are welcome to look at source code for existing implementations of a k-d tree, though of course you should not be looking at solutions to this specific CS 61B project. Make sure to cite any sources you use with the @source tag somewhere in your code. As always you should not use other student’s solutions to this project as a source, e.g. roommate’s code, something you found on pastebin, github, etc.

Proj2A Submission & Grading

Required Files for Proj2A

You are required to submit the following files:

You are also welcome to submit other files.

You will not get results from our full correctness and efficiency tests on your code when you submit on Gradescope.



We will only test your NaivePointSet for correctness, not efficiency.

K-d Tree

We will test your KDTree for correctness and efficiency. Correctness tests will be very similar to the randomized testing described in the Testing section, where we will compare your KDTree’s nearest results against our staff solution’s results (using the distance metric as described above). We will not be grading these based on efficiency, but please ensure that the method call completes in a reasonable about of time.

Efficiency tests will be similar to the correctness tests and will still require correctness, but they will also see how your solution performs with a very large number of points (hundreds of thousands) and many queries to nearest (tens of thousands). Your KDTree’s runtime will be compared to our staff KDTree’s runtime, and we will assign points based on that. If you implement the k-d tree correctly (similar to how you learn in lecture) and do not have repeated, redundant, or unnecessary function calls, you should be fine for these tests. For reference, our KDTree’s runtime while making 10,000 queries with 100,000 points is about 65-85x faster than our NaivePointSet’s runtime for the same points and queries. In addition, we will have a test that ensures your constructor is at most 10x slower than our staff KDTree’s constructor. This should not be a strict test as our constructor is relatively naive and doesn’t do anything fancier than what you learned in class.

K-d Tree Tests

You should also submit your tests, but we will not be testing your tests with an autograder.

Provided Tests

Since we are not releasing results of your submission on our full autograder, we will provide you with some very basic sanity checks to make sure your code will run with our full autograder once we run it. These will include:

These tests are by no means a good indicator of how well you will do on our full suite of tests, so make sure you write your own tests as well. Your Gradescope submission may show that you receive full points for this assignment if you pass all these tests, but this score is not your final score.

Point Distribution

The point distribution for this project will be:

Proj2B - Extrinsic MinPQ

In project 2B, you will build another data structure. Like project 2A, you will have to verify the timing and correctness of your code yourself, i.e. the autograder won’t give you very much information. Unlike 2A, where we told you exactly which algorithm to implement the steps you should take to complete it, in 2B you will have to design your own solution to the problem at hand.

A quick video overview can be found at this link. Note that this video is from Spring 2019, so some small details may be different (e.g. Extrinsic MinPQ was project 2A in spring 2019).

For this part of the project, you will build an implementation of the ExtrinsicMinPQ interface. Ultimately, this will be useful for implementing the AStarSolver class that will be described in the Proj2C spec.

If you’re not familiar with how heaps and priority queues work, please see Lecture 21 before proceeding. It is absolutely imperative that you do this before starting this assignment.

The ExtrinsicMinPQ interface quite similar to the MinPQ interface that we discussed in lecture 21. The operations are described below:

There are four key differences between this abstract data structure and the MinPQ ADT described in lecture 21.

Provided Files

We have provided a NaiveMinPQ, which is a slow but correct implementation of this interface. contains, getSmallest and removeSmallest use built-in ArrayList or Collections methods to do a brute-force search over the entire PQ. This takes time proportional to the size of the PQ, or O(n). This implementation does not throw the correct exception for the add method. This is because this exception would make this class too slow to be usable for comparing runtimes with an optimized implementation.

We have also provided a class called that can print out an array as a nice heap drawing. You’re welcome to adapt this code for your own use. It might be useful for debugging.


Your job for this part of the project is to create the ArrayHeapMinPQ class, which must implement the ExtrinsicMinPQ interface.

Your ArrayHeapMinPQ is subject to the following rules:

Note: We have not discussed how you should implement the changePriority method in lecture. You’ll have to invent this yourself. You may discuss your approach at a high level with other students (e.g. drawing out diagrams), but you should not share or look at each other’s code, nor should you work closely enough that your code may resemble each other’s.

Testing Proj2B

As in Proj2A, we will not provide any skeleton tests nor autograder messages (beyond a basic sanity check) for this project. You will be responsible for writing your own tests and ensuring the correctness of your code. You may not share tests with any other students - please ensure that all code in your possession, including tests, was written by you.

You should write your tests in a file called This file should be part of the bearmaps package.

If you’re not sure how to start writing tests, some tips follow.

  1. We encourage you to primarily write tests that evaluate correctness based on the outputs of methods that provide output (e.g. getSmallest and removeSmallest). This is in contrast to trying to directly test the states of instance variables (see tip #3 below). For example, to test changePriority, you might use a sequence of add operations, a changePriority call, and then finally check the output of a removeSmallest call. This is (usually) a better idea than iterating over your private variables to see if changePriority is setting some specific instance variable.

  2. Write tests for functions in the order that you write them. You might even find it helpful to write the tests first. Since each function may call previously written functions, this helps ensure that you are building a solid foundation.

  3. If you want to write tests that require looking at your private instance variables, these tests should be part of the ArrayHeapMinPQ class itself. For example, if you want to write a test that only calls the add method, there’s no way to write it in the manner suggested in tip #1 in ArrayHeapMinPQTest. Suppose that you want to verify that your array is [1, 2, 4, 5, 3] after inserting 5, 4, 3, 2, 1. In this case, a hypothetical testAdd54321 method would need to be part of the ArrayHeapMinPQ class since the ArrayHeapMinPQTest should not be able to access the private instance variables. It’s debatable whether you should even write such tests, since tests of type #1 should ideally catch any errors. Follow your heart.

  4. One annoying issue in Java is testing of private helper methods. For example, suppose you have a leftChild method in ArrayHeapMinPQ that you’d like to test in ArrayHeapMinPQTest. If this method is private, then the test file will be unable to call the method. For this project, if you have helper methods you’d like to test, you can either include the tests in ArrayHeapMinPQ, or you should make those helper methods “package private”. To do this, simply remove the access modifier completely. That is, rather than saying public or private, you should add no access modifier at all. This will make this method accessible only to other classes in the package. You can think of package private as a level of access control in between public (anybody can use it) and private (only this class can use it).

  5. Don’t forget edge cases - consider how the heap could look before inserting or removing items, and ensure that your code handles all possible cases (for example, sinking a node when its priority is greater than both of its children).

  6. Rather than thinking about all possible edge cases, feel free to perform randomized tests by comparing the outputs of your data structure and the provided NaiveMinPQ, similar to what you did in Project 2A.

  7. To test the runtime of your code, you should use the System.currentTimeMillis method or the edu.princeton.cs.algs4.Stopwatch class, similarly to what you did in Lab 5 and Project 2A. Keep in mind that calling the add method N times will take O(Nlog(N)) total time if each add call is O(log(N)) time. Also, keep in mind that if you call the contains method N times on a data structure of size N, you’d again expect total runtime to grow as O(Nlog(N)). Note: If you attempt to fit a log function using some sort of fancy software, you will probably be disappointed. Timing tests are likely to be too noisy for such fitting tools to work well. If you’re expect O(Nlog(N)) and you’re getting something that looks linear, then you’re in good shape.

  8. To test that your code throws the proper exceptions, you should leverage JUnit’s power. Read this post to see how to do that. Note that as soon as your code throws some type of Exception, it will not complete the rest of the lines in that testcase. So, if you have multiple cases that should each throw an exception, you should write multiple testcases: one for each case.

External Resources

You are welcome to look at source code for existing priority queue implementations. Make sure to cite any sources you use with the @source tag somewhere in your code. As always you should not use other student’s solutions to this project as a source, e.g. roommate’s code, something you found on pastebin, github, etc.

You might find the MinPQ implementation from the optional textbook to be a helpful reference. However, you should not copy and paste into your implementation and then try to figure out how to bend it to match our spec. While it might seem like this will save you time and effort, it will end up generating more work than just building everything yourself from scratch. Many students tried doing something similar with from lecture and from Project 1A, and the results were generally very bad.

Project 2A FAQ

Q: How do I create a Node[] if my Node class stores objects of generic type T? If I try (Node[]) new Object[10], I get a classCastException at runtime, but if I try new Node[10], I get a generic array creation error at compile time. This is one particularly annoying issue with Java generics. Without going into the details, the easy fix is that you can simply say new ArrayHeapMinPQ.Node[10] and it will work. See this stack overflow post for more details.

Q: When I make a test for changePriority and test NaiveMinPQ against ArrayHeapMinPQ, the runtime for ArrayHeapMinPQ isn’t substantially better. Let’s consider how NaiveMinPQ stores items. It stores them in a list, left to right. So if I insert items 0-1,000,000, it would have item 0 at the 0th index, 1 at the 1st index, and so on. Now, when it changes the priority of an item, it has to scan this list from left to right looking for this item. If my test were to insert items 0-1,000,000, then change the priority of items 0-1,000, this is actually the best case input for NaiveMinPQ::changePriority, since these are the closest items it can find. We recommend changing your test to randomly select 1000 items in the heap and change their priority.

Project 2B Submission & Grading

Required Files for Proj2A

Your Project 2B submission must contain the following files:

You can also include other files that you create, e.g.



We will test your ArrayMinHeapPQ for correctness and efficiency. Correctness tests will be very similar to the randomized testing described in the Randomized Testing section of the Proj2A spec above, where we will compare your ArrayMinHeapPQ’s getSmallest, removeSmallest, add, and changePriority results against our staff solution’s results.

Efficiency tests will be similar to the correctness tests and will still require correctness, but they will also see how your solution performs with a very large number of items (millions) and many queries to changePriority (thousands). Your changePriority runtime will be compared to the staff solution’s changePriority runtime, and we will assign points based on that. If you implement changePriority correctly (O(log(n))) and do not have repeated, redundant, or unnecessary function calls, your runtime should be fine for these tests.


You must submit your tests. However, unlike Project 1B, your tests will not be tested with an autograder.

Point Distribution