Java - DukeJava - 4-3-4 Programming Exercise 4 WordGram Class

DukeJava - 4-3-4 Programming Exercise 4 WordGram Class

[toc]

Java-Programming-and-Software-Engineering-Fundamentals-Specialization

• 4.Java-Programming-Principles-of-Software-Design
  • N-Grams: Predictive Text
    • 4-3-1 Programming Exercise 1 Generating Random Text
    • 4-3-2 Programming Exercise 2 Interface-and-Abstract-Class
    • 4-3-3 Programming Exercise 3 Word N-Grams
    • 4-3-4 Programming Exercise 4 WordGram Class

Resource Link: https://www.dukelearntoprogram.com/course4/index.php ProjectCode: https://github.com/ocholuo/language/tree/master/0.project/javademo

---

The class WordGram is a class that represents words that have an ordering. For example, a WordGram might be the four words “this” “is” “a” test”, in this order. The WordGram class has two private variables, a String array myWords to store the words in order, one word per slot, and a private integer myHash you will use to be able to use WordGrams as a key with a HashMap. This class has several methods
The constructor has three parameters, a String array named source, an integer named start, and an integer named size. The constructor copies the size number of words from source starting at the position start into a new WordGram.
The method wordAt has one integer argument name index. This method returns the word in the WordGram at position index.


The class WordGramTester has methods for testing the WordGram. You may find these helpful in testing the methods you write.
The void method testWordGram builds and prints several WordGrams from a String.
The void method testWordGramEquals tests if two WordGrams are equivalent.

---

Assignment 1: Complete WordGram

Write the method length that has no parameters and returns the length of the WordGram. This method has been started for you.

The method toString that has no parameters. It prints a WordGram out, showing all the words in the WordGram on one line separated by spaces. This method has been started for you.

Write the method equals that has one parameter of type Object named o. This method returns true if two WordGrams are equal and false otherwise. This method has been started for you.

Write the method shiftAdd that has one String parameter word. This method should return a new WordGram the same size as the original, consisting of each word shifted down one index (for example the word in slot 1 would move to slot 0, the word in slot 2 would move to slot 1, etc.) and word added to the end of the new WordGram. Be sure to test this method. For example, if a WordGram of size 4 is “this” “is” “a” “test”, and shiftAdd is called with the argument “yes”, then the method would return a new WordGram ”is” “a” “test” “yes”. This method should not alter the WordGram on which it is called.

---

Assignment 2: MarkovWord with WordGram

Create the MarkovWord class that implements IMarkovModel. This class should have three private variables, a String array named myText, a Random variable named myRandom, and an integer variable named myOrder. This class should have the following methods, similar to what the MarkovWordOne class had, but extended for handling a larger number of words. You should copy the body of MarkovWordOne and then modify it. The methods in MarkovWord are:

A constructor with one integer parameter that is the order (how many words to use in prediction). This method should initialize myOrder and myRandom.

The void method setRandom has one integer parameter named seed. Using this method will allow you to generate the same random text each time, which will help in testing your program. This method should be the same as it was in MarkovWordOne.

The void method setTraining has one String parameter named text. The String text is split into words and stored in myText. The words are used to initialize the training text. This method should be the same as it was in MarkovWordOne.

The indexOf method has three parameters, a String array of all the words in the training text named words, a WordGram named target, and an integer named start indicating where to start looking for a WordGram match in words. This method should return the first position from start that has words in the array words that match the WordGram target. If there is no such match then return -1.

The getFollows method has one WordGram parameter named kGram. This method returns an ArrayList of all the single words that immediately follow an instance of the WordGram parameter somewhere in the training text. This method should call indexOf to find these matches.

The getRandomText method has one integer parameter named numWords. This method generates and returns random text that has numWords words. This class generates each word by randomly choosing a word from the training text that follows the current word(s) in the training text. When you copied the body of MarkovWordOne into the MarkovWord class, you copied this method from MarkovWordOne. Much of the code from the copied method will still be correct for MarkovWord, but you will need to make a few changes so that it works for any order (not just order one), and uses WordGram objects. You may want to use the shiftAdd method you wrote in WordGram.

---

Assignment 3: EfficientMarkovWord with WordGram

add a method named hashCode that has no parameters. This method should return an integer that is a hash code that represents the WordGram. Note that String has a method hashCode. You may want to create a String from the WordGram and use the String hashCode method.

Write a new class named EfficientMarkovWord (make a copy of MarkovWord to start with) that implements IMarkovModel and that builds a HashMap to calculate the follows ArrayList for each possible WordGram only once, and then uses the HashMap to look at the list of characters following when it is needed. This class should include:

a method named buildMap to build the HashMap (Be sure to handle the case at the end where there is not a follow character. If that WordGram is not in the HashMap yet, then it should be put in mapped to an empty ArrayList. If that key is already in the HashMap, then do not enter anything for this case.)

a getFollows method, but this getFollows method should be much shorter, as it can look up the WordGram, instead of computing it each time.

To test your HashMap to make sure it is built correctly, write the void method printHashMapInfo in the EfficientMarkovWord class. This method should print out the following information about the HashMap:
- Print the HashMap (all the keys and their corresponding values). Only do this if the HashMap is small.
- Print the number of keys in the HashMap
- Print the size of the largest value in the HashMap—that is, the size of the largest ArrayList of characters
- Print the keys that have the maximum size value.

create an order-2 EfficientMarkovWord with
- seed 42
- the training text is “this is a test yes this is really a test”
- the size of the text generated is 50

leave the order at 2, the random seed at 42, and the size of the text generated at 50. Set the training text to the following String
- It has 10 keys in the HashMap
- The maximum number of elements following a key is 3
- There are two WordGrams that each have three follow items. The key “this” “is” has the follow words are “a”, “really” and “wow”. The key “a” “test” has the follow words “yes”, “yes”, and “this”.
- Note the follow of “is” “wow” should be [ ], an empty array.

create a void method named compareMethods that runs a MarkovWord and an EfficientMarkovWord
- Make both order-2 Markov models
- Use seed 42 and set the length of text to generate to be 100
- Both should call runModel that generates random text three times for each.

Run the MarkovWord first and then the EfficientMarkovWord with the file “hawthorne.txt” as the training text. One of them should be noticeably faster. You can calculate the time each takes by using System.nanoTime() for the current time.

3312586086453 : MarkovWord

32022293536 : EfficientMarkovWord

.