// APS105 Lab6 11/05/06
// Geoffrey Lang
// 995155772
// langgeof@ecf.utoronto.ca

#include <iostream>
#include <fstream>

using namespace std;

const int MAX_WORDS = 10000;
const int MAX_WORD_LENGTH = 25;

bool goodLetter(char a);
void printArray(char words[][MAX_WORD_LENGTH], int n);
void selectionSort(char words[][MAX_WORD_LENGTH], int n);
void deleteDuplicates(char words[][MAX_WORD_LENGTH], int& n);

int main (int argc, char *argv[])
{
        char current;
        int i = 0, j = 0;
        bool twolines = 0;

        char words[MAX_WORDS][MAX_WORD_LENGTH];
        char temp[MAX_WORD_LENGTH];


        ifstream in_stream;                                     // Opens the file written after the
        in_stream.open(argv[1]);                                // name of this program when it is run

        while(!in_stream.eof())
        {
                in_stream.get(current);
                current = tolower(current);

                if((current == ' ' || current == '\n') && twolines == 1)// Move to next line if there is
                {                                                       // a space in the input
                        while(!goodLetter(words[i][strlen(words[i])-1]))
                        {
                                words[i][strlen(words[i])-1] = '\0';
                        }
                        i++;
                        j = 0;
                        twolines = 0;
                }

                if(j == 0 && goodLetter(current) || j != 0)     // Check that the first letter isn't punctuation
                {                                               // before putting the word into the array and
                        words[i][j] = current;                  // if its not the first letter, put it in anyway
                        j++;
                        twolines = 1;
                }
        }

        selectionSort(words,i);                                 // Apply selection sort to the array
                                                                // and then print it (for testing
//      printArray(words,i);                                    // purposes only)
        deleteDuplicates(words,i);                              // Delete the duplicates in the array
                                                                // and print it to check that they are
        printArray(words,i);                                    // all deleted

        return 0;
}

bool goodLetter(char a)                                         // Function to check if a character is
{                                                               // a character that can start or end
        if((a >= 'a' && a <= 'z') || (a >= '0' && a <= '9'))    // a word--must be a letter or a
        {                                                       // number. Returns 1 if the character
                return 1;                                       // is a letter or number, 0 if not
        }
        else
        {
                return 0;
        }
}

void printArray(char words[][MAX_WORD_LENGTH], int n)           // Function to print an array, one
{                                                               // word per line
        for(int x = 0; x < n; x++)
        {
                for (int y = 0; words[x][y] != '\0'; y++)
                {
                        cout << words[x][y];
                }
                cout << endl;
        }
        return;
}

void selectionSort(char words[][MAX_WORD_LENGTH], int n)        // Function to sort an array of
{                                                               // c-strings in alphabetical
        for(int i = 0; i < n - 1; i++)                          // order
        {
                int smallest = i;
                for(int j = i+1;j < n; j++)
                {
                        if(strcmp(words[j],words[smallest]) < 0)
                        {
                                smallest = j;
                        }
                }
                char temp[1][MAX_WORD_LENGTH];
                strncpy(temp[0],words[i],MAX_WORD_LENGTH);
                strncpy(words[i],words[smallest],MAX_WORD_LENGTH);
                strncpy(words[smallest],temp[0],MAX_WORD_LENGTH);
        }
        return;
}

void deleteDuplicates(char words[][MAX_WORD_LENGTH], int& n)            // Function to delete all duplicate
{                                                                       // enteries in an array of sorted
        int i=0, j=0, removed=0;                                        // c-strings
        while(i < n - 1 )
        {
                while (strncmp(words[i-j],words[i+1],MAX_WORD_LENGTH) == 0)
                {
                        j++;
                        i++;
                }

                if (j > 0)
                {
                        strncpy(words[i-j+1],words[i+1],MAX_WORD_LENGTH);
                        j--;
                }
                removed++;
        }
        n = removed;
        return;
}