For this week's lab, you will use two of the classes in the Java Collection Framework: HashSet and TreeSet. You will use these classes to implement a spell checker.
This lab is due at the time of the next lab, next Tuesday, March 31. Please copy your work into the homework directory by that time.
Your work from labs 6, 7, and 8 is due this week. I will not look for this work until this weekend. Make sure that it is done before Saturday morning. You should have a working genetic programming project, including the work that you did for both Lab 6 and Lab 7. I will look for this work in the CVS repository that you created for Lab 8. You should also have a written report on your work with the debugger in Lab 8, which you can turn in in class (or let me know exactly where I can find it).
For this lab, you will need to use some of the methods that are defined in the Set interface. Recall that if set is a Set, then the following methods are defined:
You will also need to be able to traverse a set, using either an iterator or a for-each loop.
The file /classes/s09/cs225/words.txt contains a list of English words, with one word on each line. You will look up words in this list to check whether they are correctly spelled. To make the list easy to use, you can store the words in a set. Since there is no need to have the words stored in order, you can use a HashSet for maximum efficiency.
Use a Scanner to read the file. Recall that you can create a scanner, filein, for reading from a file with a statement such as:
filein = new Scanner(new File("/classes/s09/cs225/words.txt"));
and that a file can be processed, token by token, in a loop such as:
while (filein.hasNext()) {
String tk = filein.next();
process(tk); // do something with the token
}
(For the wordlist file, a token is simply a word.)
Start your main program by reading the words from /classes/s09/cs225/words.txt and storing them in a HashSet<String>. For the purposes of this program, convert all words to lower case before putting them in the set. To make sure that you've read all the words, check the size of the set. (It should be 73845.) You could also use the contains method to check for the presence of some common word in the set.
Once you have the list of words in a set, it's easy to read the words from a file and check whether each word is in the set. Start by letting the user select a file. You can either let the user type the name of the file or you can use the following method:
/**
* Lets the user select an input file using a standard file
* selection dialog box. If the user cancels the dialog without
* selecting a file, the return value is null.
*/
static File getInputFileNameFromUser() {
JFileChooser fileDialog = new JFileChooser();
fileDialog.setDialogTitle("Select File for Input");
int option = fileDialog.showOpenDialog(null);
if (option != JFileChooser.APPROVE_OPTION)
return null;
else
return fileDialog.getSelectedFile();
}
Use a Scanner to read the words from the selected file. In order to skip over any non-letter characters in the file, you can use the following command just after creating the scanner (where in is the variable name for the scanner):
in.useDelimiter("[^a-zA-Z]+");
(In this statement, "[^a-zA-Z]+" is a regular expression that matches any sequence of one or more non-letter characters. This essentially makes the scanner treat any non-letter the way it would ordinarily treat a space.)
You can then go through the file, read each word (converting it to lower case) and check whether the set contains the word. At this point, just print out any word that you find that is not in the dictionary.
A spell checker shouldn't just tell you what words are misspelled -- it should also give you a list of possible correct spellings for that word. Write a method
static TreeSet corrections(String badWord, HashSet dictionary)
that creates and returns a TreeSet<String> containing variations on badWord that are contained in the dictionary. In your main program, when you find a word that is not in the set of legal words, pass that word to this method (along with the set). Take the return value and output any words that it contains; these are the suggested correct spellings of the misspelled word. Here for example is part of the output from my program when it was run with the HTML source of this page as input:
html: (no suggestions) cpsc: (no suggestions) hashset: hash set treeset: tree set cvs: cs, vs isempty: is empty href: ref txt: tat, tet, text, tit, tot, tut filein: file in pre: are, ere, ire, ore, pare, pee, per, pie, poe, pore, prep, pres, prey, pro, pry, pure, pyre, re hasnext: has next wordlist: word list getinputfilenamefromuser: (no suggestions) jfilechooser: (no suggestions) filedialog: file dialog setdialogtitle: (no suggestions) int: ant, dint, hint, in, ina, inc, ind, ink, inn, ins, inti, into, it, lint, mint, nit, pint, tint
Note that I have written my program so that it will not output the same misspelled word more than once. (I do this by keeping a set of misspelled words that have been output.) If my corrections() method returns an empty set, I output the message "(no suggestions)". Since the corrections are stored in a tree set, they are automatically printed out in alphabetical order with no repeats.
The possible corrections that I consider are as follows:
For constructing the possible corrections, you will have to make extensive use of substrings. If w is a string, then w.substring(0,i) is the string consisting of the first i characters in w (not including the character in position i, which would be character number i+1). And w.substring(i) consists of the characters of w from position i through the end of the string. For example, if ch is a character, then you can change the i-th character of w to ch with the statement:
String s = w.substring(0,i) + ch + w.substring(i+1);
Also, you will find it convenient to use a for loop in which the loop control variable is a char:
for (char ch = 'a'; ch <= 'z'; ch++) { ...