Description
CS2103 Introduction In this project you will develop software instructure to model large graphs along with a search engine to find shortest paths between any pair of nodes in a graph. As an example application illustrating how to use this infrastructure, you will develop a search engine for the Internet Movie Database (IMDB) that allows users to find a shortest path of actor nodes and movie nodes that connect any two specified actors (if such a path exists). The code that your team will write to load data and find paths will interface with code (written by me) that implements the graphical user interface (GUI — see figure below) of the application. Parsing the IMDB data One of the tedious but crucial aspects of real-world computer programming is converting data from one format to another. It’s important to become efficient and very reliable at doing this, and this project provides a practice opportunity. In particular, you will create a class IMDBGraphImpl that implements the IMDBGraph interface. Implementing this class will require you to parse data files that contain information on actresses/actors and the movies in which they starred. While you are welcome to use more sophisticated parsing methods such as regular expressions, these are not required for this assignment. In fact, my own implementation simply uses a Scanner (just its hasNextLine and nextLine methods — nothing fancier) as well as some standard String methods: indexOf, substring, and lastIndexOf. Project 3 https://ia.wpi.edu/cs2103/resources.php?page=show_project&id=12 1 of 5 11/11/2018, 6:08 PM Data format IMDB uses a non-standard and rather unfortunate data format…as sometimes happens in the real world. The dataset consists of two data files: actors.list and actresses.list. Each file starts with a preamble that describes the file structure. The actors.list data file begins in earnest with the lines: THE ACTORS LIST =============== Name Titles —- —— (The actresses.list begins analogously.) In each file, actresses/actors are listed alphabetically. For each actress/actor, the list of movies in which she/he starred is listed, followed by an empty line to separate the current actress/actor from the next actress/actor, e.g.: $hutter Battle of the Sexes (2017) (as $hutter Boy) [Bobby Riggs Fan] <10> NVTION: The Star Nation Rapumentary (2016) (as $hutter Boy) [Himself] <1> Secret in Their Eyes (2015) (uncredited) [2002 Dodger Fan] Steve Jobs (2015) (uncredited) [1988 Opera House Patron] Straight Outta Compton (2015) (uncredited) [Club Patron/Dopeman] In the example above, an actor named $hutter (for some reason he spells his name with a $ sign) has starred in 5 movies and/or television shows, one movie/show shown on each line. At least one tab character (\t) separates the actor from the first movie/show, and at least one tab character precedes each movie/show given on subsequent lines. Note that the actress’/actor’s name is not repeated for each movie/show in which she/he starred — it is shown only once for the first movie/show. Along with the name of the movie/show and the year, additional information may be given including the name of the character and the billing position of the actor in the credits. In this programming project, you should only parse the title and year of each movie. In the example above, your parsing code should thus produce the following five strings for the actor $hutter: Battle of the Sexes (2017) NVTION: The Star Nation Rapumentary (2016) Secret in Their Eyes (2015) Steve Jobs (2015) Straight Outta Compton (2015) In particular, the string representing the title of each movie that your code parses should include the year in parentheses as part of the title string. Parsing the data The actors.list and actresses.list data files are not in ASCII format; rather, they are in ISO-8859-1 format. This file format allows the use of non-ASCII characters that are common in many languages (but not English). It is essential that you use the proper character decoder when parsing the files. We recommend that you use the java.util.Scanner class, e.g.: final Scanner scanner = new Scanner(new File(actorsFilename), “ISO-8859-1”); The line of code above instantiates a new Scanner object that reads from a java.io.File (on disk) and decodes it using a ISO-8859-1 decoder. From then on, you can use all the standard methods available through the Scanner class; some of the more useful ones include: hasNextLine nextLine, and findInLine. Keeping the memory usage tractable The IMDB contains about 30,000,000 lines of data that your code must parse and analyze in order to build the graph. In order to reduce the space and time costs involved in this project, you should exclude any TV series as well as any TV movie. As described in the data files, TV movies contain (TV) in the line describing them, e.g.: Sweet Talkin’ Guys (1991) (TV) (archive footage) [Himself] TV series have quotation marks around the titles, e.g.: “All You Need Is Love” (1977) {Introduction (#1.1)} [Himself] In addition, you should also exclude any actress or actor who starred only in TV series/TV movies; an example of such an Project 3 https://ia.wpi.edu/cs2103/resources.php?page=show_project&id=12 2 of 5 11/11/2018, 6:08 PM actress/actor is Aaltio, Mikko: Aaltio, Mikko “Bitwisards” (2016) {Bitwisardin saari (#1.10)} [Jorma] <6> “Bitwisards” (2016) {Pitkästä ilosta ja tyhjännauramisesta (#1.11)} [Jorma] <10> “Bitwisards” (2016) {Uusi aika (#1.8)} [Jorma] <10> because he starred only in shows in a TV series. Note: do not worry about trying to correct any typos that may be present in IMDB, or trying to “merge” two actors whose names differ in trivial ways (e.g., the presence/absence of a middle initial). You should assume that each actress/actor and each movie title in the actors.list and actresses.list datasets are fully correct (even though this is not actually the case). Implementing the IMDBGraph interface As mentioned above, one of your tasks in this project is to implement the IMDBGraph interface. This class should contain a public constructor that takes two parameters (in order): 1. The absolute path (e.g., “C:\MyDirectory\actors.list”) of the actors.list file. 2. The absolute path (e.g., “C:\MyDirectory\actresses.list”) of the actresses.list file. Your code must load the IMDB data from the files whose absolute paths are specified in the parameters passed to the constructors. In addition, these constructors must throw a java.io.IOException — they should not attempt to catch this exception themselves. Both of these criteria are essential to enable us to grade your code automatically. Specifically, your code should contain the following constructor: public IMDBGraphImpl (String actorsFilename, String actressesFilename) throws IOException { // Load data from the specified actorsFilename and actressesFilename … } Correspondingly, our test code will invoke your constructors as: graph = new IMDBGraphImpl(“IMDB/actors.list”, “IMDB/actresses.list”); Finding shortest paths between nodes in a graph To find a shortest path between node s and t in a graph, you should implement a breadth-first search (BFS), as described during class. Once you have found a shortest path — if one exists — you then need to backtrack from t back to s and record the sequence of nodes that were traversed. The result should then be returned back to the caller (the GraphSearchGUI, in this case). Even for large social network graphs consisting of millions of nodes, BFS will be very fast, as it operates in O(n) time (average case), where n is the number of nodes in the graph. For this part of the assignment, you should create a class called GraphSearchEngineImpl that implements the GraphSearchEngine interface. The findShortestPath method of your GraphSearchEngineImpl should return an instance of type List in which the first element of the list is s, the last element of the list is t, and the intermediate nodes constitute a shortest path (alternating between movies and actresses/actors) that connect nodes s and t. If no shortest path exists between the pair of nodes, then your method should return null. Make sure that your GraphSearchEngineImpl is not “tied” to the IMDB data in any way — the search engine should be useful for any graph of Node objects. Requirements R1: Create a class called IMDBGraphImplthat implements the Graph interface: The class should load data from the two files whose filenames are specified in the constructor. Your parsing code should exclude TV series, TV movies, and any actresses/actors who starred only in TV series/movies. The class should have a public constructor, which can throw an IOException, that take two parameters specifying the absolute paths of the actors and actresses data files, in that order. The class should construct a graph based on the data parsed from these files. To make sure that everyone is parsing the same data files, you must use the IMDB data contained in the Zip file at the following link: IMDB.zip. 1. 2. R2: Implement BFS within a class GraphSearchEngineImpl that implements GraphSearchEngine. Your search engine Project 3 https://ia.wpi.edu/cs2103/resources.php?page=show_project&id=12 3 of 5 11/11/2018, 6:08 PM should be able to find shortest paths between any pair of Node objects (or return null if no such path exists) — whether they are IMDB nodes, CiteSeer (scientific publication) nodes, or anything else. Make sure that the files you submit are named exactly as described above so that our automatic test code can give you credit for your work. Note that you may not change any of the interfaces that we provide in any way. Teamwork You may work as a team on this project; the maximum team size is 2. Design and Style Your code must adhere to reasonable Java style. In particular, please adhere to the following guidelines: Each class name should be a singular noun that can be easily pluralized. Class names should be in CamelCase; variables should be in mixedCase. Avoid “magic numbers” in your code (e.g., for (int i = 0; i < 999 /*magic number*/; i++)). Instead, use constants, e.g., private static final int NUM_ELEPHANTS_IN_THE_ROOM = 999;, defined at the top of your class file. Use whitespace consistently. No method should exceed 50 lines of code (for a “reasonable” maximum line length, e.g., 100 characters). If your method is larger than that, it’s probably a sign it should be decomposed into a few helper methods. Use comments to explain non-trivial aspects of code. Use a Javadoc comment to explain what each method does, what parameters it takes, and what it returns. Use the /**…*/ syntax along with @param and @return tags, as appropriate. Use the final keyword whenever possible. Use the most restrictive access modifiers (e.g., private, default, protected>, public), for both variables and methods, that you can. Note that this does not mean you can never use non-private access; it just means you should have a good reason for doing so. Declare variables using the weakest type (e.g., an interface rather than a specific class implementation) you can; ithen instantiate new objects according to the actual class you need. This will help to ensure maximum flexibility of your code. For example, instead of final ArrayList list = new ArrayList(); use final List list = new ArrayList(); If, on the other hand, you have a good reason for using the actual type of the object you instantiate (e.g., you need to access specific methods of ArrayList that are not part of the List interface), then it’s fine to declare the variable with a stronger type. Getting started Download the Project3 starter file. Download the IMDB data. Note: the GraphSearchGUI class is the motivating application of this project, but you are not required to use it or extend it in any way. It’s mainly just for fun. Testing In this project, we will not grade your test code. However, we highly recommend that you create a toy dataset to help you develop and debug your BFS code. Creating toy datasets is a very useful technique in a variety of software engineering (and computer science) domains. This will greatly reduce the amount of computation that is required to make progress on the project. It will also make it easier to understand and eliminate any bugs that your code might contain. What to Submit Project 3 https://ia.wpi.edu/cs2103/resources.php?page=show_project&id=12 4 of 5 11/11/2018, 6:08 PM Create a Zip file containing IMDBGraphImpl.java and GraphSearchEngineImpl.java, along with any other files that you wish. Submit the Zip file you created to Canvas. Your code will be graded by automatic scripts. Submission deadline: Wednesday, November 14, at 11:59pm EDT. Project 3 https://ia.wpi.edu/cs2103/resources.php?page=show_project&id=12 5 of 5 11/11/2018, 6:08 PM