Difference between revisions of "CSC212 Homework 8 2014"

From dftwiki3
Jump to: navigation, search
(Submission)
(Same Number of Followers)
Line 68: Line 68:
 
===Same Number of Followers===
 
===Same Number of Followers===
 
<br />
 
<br />
* If there tweets from people who have the same largest number of followers, then your program should output '''all''' of them.
+
* If the file contains tweets from people who have the same largest number of followers, then your program should output '''all''' of the tweets.
 
<br />
 
<br />
  

Revision as of 08:03, 16 November 2014

--D. Thiebaut (talk) 20:30, 12 November 2014 (EST)




This homework assignment is due Nov. 21st, 2014, at 11:55 p.m.




Problem #1


  • Write a program called Hw8_1.java based on the quicksort program of Lab 12, and that contains a new method called quickQuicksort(... ), that has the following properties:
  • It receives an array A of ints (int[] A), along with two ints defining the low and high indexes it must apply itself.
  • It receives an integer N, with N' less than or equal to the size of A.


        public static void quickQuicksort( int[] A,int low, int high, int N ) {
            ...
        }


  • It partially sorts the array A so that the lowest N integers of the array end up sorted in the left-most side of the array (the part of the array that contains all the cells with low indexes). For example, assume that quickQuicksort receives A = [10, 3, 5, 1, 2, 9, 0, 3], and N = 3. Quicksort will stop as soon as the lowest 3 integers of A are located in the cells of A with lowest index, i.e. A = [0, 1, 2, 5, 10, 3, 9]. Note that only the lowest 3 integers are sorted in the left-most part of the array. The remaining part of A is unsorted (but can be partially sorted).
  • The purpose of quickQuicksort is to have a very fast algorithm for locating the N lowest (or N highest) items of a large collection, without having to spend the time to sort the whole array. It is an alternative solution to the problem we solved in class with a heap.


Performance


  • Your program has to be robust and avoid O(N2) performance on sorted arrays.
  • Your program's execution time will be measured and compared to that of the solution program, on different arrays, some random, some sorted in increasing order, and others sorted in decreasing order. Your function will also be tested with values of N ranging from 0 to the length of A.
  • The grade will be proportional to the speed of your program.


Submission


  • Submit your program to Moodle, in the Homework 8, Problem 1 section.
  • The test program will run your program and the solution program several times on large (500,000 items) arrays of sorted and unsorted ints. Each array tested is given to both the solution program and your program, to make the comparison fair. The average execution times for both programs are computed. A grade of 100 is generated if the average execution time of your program is no more than 110% of the execution time of the solution program. The grade is then decreased the longer your program takes compared to the solution program.
  • Testing your program this way may take a minute or longer...


Problem #2

Tweets.jpg


Your Assignment


  • Write a Java program that reads a collection of tweets stored in a csv file, and outputs the tweet corresponding to the person with the most followers.
  • Your program should be called Hw8_2.java.


Requirements


Imports


  • Challenge: You are to use only the following libraries in your program:


import java.io.File;
import java.io.FileNotFoundException;
import java.net.URL;
import java.util.PriorityQueue;
import java.util.Scanner;


File Name passed on the Command Line


  • Your program should get the name of the file of tweets on the command line. In Eclipse, this means providing the name in the Run Configurations/Arguments/Program Arguments. On Beowulf2, this means providing the name on the command line, as in java Hw8_2 tweets.csv.


Same Number of Followers


  • If the file contains tweets from people who have the same largest number of followers, then your program should output all of the tweets.


Additional Information


CSV File


  • You can use this file as a practice file for your Java program: tweets.csv
  • To make your csv file accessible to your program, you have two options:
In Eclipse: click on the src folder containing your HW8_2.java file, and right/control click to add a New File. Call it tweets.csv.txt. Note the .txt extension; if you do not have this extension, Eclipse might switch to a spreadsheet editor when you open the file. The .txt extension will fool Eclipse in opening a text editor.
Copy/paste the sample tweets from the link above. Once you have saved the tweets in your tweets.csv.txt file, refactor/rename the file and remove the ".txt" extension, to make the name tweets.csv.
To access the file, use this code as inspiration:


String fileName = "tweets.csv";
URL url = Hw8_2.class.getClassLoader().getResource( fileName );
Scanner s = new Scanner( new File( url.getPath() ) );


On Beowulf2: put the file in the same directory where your program resides, and your code will find the file without problem.


Expected Output


  • Here are some examples of how your program should behave. The program is run on Beowulf2 in this example. Note that the testing software will expect your program to output the same information the solution program outputs for missing file name on the command line, or for files not found conditions.


java Hw8_2 tweets.csv 
11/14/14 18:07: https://twitter.com/karessaNV_M/status/533237307751792640 (28568 followers) (follows 23674) At least you don't h...

java Hw8_2
Syntax: java Hw8_2 filename.csv
		where filename.csv is a csv file in the current path

java Hw8_2 badname.csv
Cannot find specified file badname.csv


  • Note that the program truncates the text of the tweets and adds the ellipses (...).
    The code looks like this: TweetText.substring(0, 20)+"..."



Submission


  • Submit your program on Moodle, in the Homework 8, Problem 2 section.
  • Your program will be tested with several different csv files, with different properties. Try to anticipate ways in which these files will be created.