/*  12.010 Homework #3 Question 2
   Program to read a paragraph and comoute the mean and rms about
   the mean of the number of characters per word and the number of 
   words per sentence.
*/

#include <stdio.h>
#include <string.h>
#include <ctype.h>
#include <math.h>

int main(argc, argv) 
int argc ; // Number of arguments
char  *argv[] ; // Pointer to arguments 

{

int  inc ;     // Input Character read from file (EOF means end of file)
int  num_char = 0 ; //  Number of characters in current word
int  num_word = 0 ; //  Number of words in while file
int  num_sent = 0 ; //  Number of sentences in file
int  num_word_insent = 0 ;  //  Number of words in current sentence
int  sum_num_char = 0 ; //  Sum of total number of chacters
int  sum_num_word = 0 ; //  Sum of number of words
int  sum_num_char_sq = 0 ; //  Sum of number of characters squared
int  sum_num_word_sq = 0 ; //  Sun of number of words squared

int still_in_word = 0; /*  set non-zero for hypenated word to show we
                        are still in word on new line */
int in_gap = 0 ;  /*  Set non-zero when space is found incase there are
                  more than one-space between words and when 
                  the end of sentence is found. */

float mean_char, rms_char  ; // Mean and RMS number of characters per word
float mean_sent, rms_sent  ; // Mean and RMS of words per sentence.
float rmssq ; // Square of RMS
   
char filename[128];  /* Name of file to read, passes in runstring */

FILE *fp, *fopen() ; /* File pointer to open files

/* See how many arguments were passed */
   if ( argc > 1 ) {
       strcpy(filename,argv[1]);
   }
   else {
       strcpy(filename,"Q2_text.txt");
   }
   printf("\n12.010 HW 3 Q2:\nReading file %s\n",filename);

/* Now try to open the file */
   fp = fopen(filename,"r");
   if ( fp == NULL ) {
       printf("Error opening file %s\n",filename);
       return(-1);
   }

/* Start reading file.  We will read one-character at a time
   and count the statistics as we go.  We could also read 
   line-by-line as in the fortran code. */

  inc = 0;
  for ( ; inc != EOF ; ) {
     inc = fgetc(fp);
     if ( inc != EOF && inc != '\n' ) {

        /* See how we should count this character */
        if( (inc >= 'a' && inc <= 'z') || (inc >= 'A' && inc <= 'Z') ) {
            /* regulare letter */
            num_char++ ; in_gap = 0;
        }
        else if( inc == ' '  && in_gap == 0 ) {
           /* End of a word or line but not repeated space */
            num_word++ ; num_word_insent++ ;
            sum_num_char = sum_num_char + num_char ;
            sum_num_char_sq = sum_num_char_sq + num_char*num_char; 
          //  Reset number of characters and that we are in gap
            num_char = 0 ;
            in_gap = 1;
        }
        else if( inc == '.' || inc == '?' ) {
           /* End of a sentencc and so increment same as space */
           if( num_char > 0 ) {
               num_word++ ;
               num_word_insent++ ;
               sum_num_char = sum_num_char + num_char ;
               sum_num_char_sq = sum_num_char_sq + num_char*num_char ; 
          //   Reset number of characters and that we are in gap
               num_char = 0 ; 
          } 
          /* Now process end of sentence */
               num_sent++ ; 
               sum_num_word = sum_num_word + num_word_insent ;
               sum_num_word_sq = sum_num_word_sq + num_word_insent*num_word_insent;
               num_word_insent = 0 ;
               in_gap = 1; 
        }
        else if ( inc == '-' ) {
            /*     Hypenated word, so keep character count going */
            still_in_word = 1;
            /*   Ignore any other characeters (:, ;, numbers) */
        } 
     }  /* End of not EOF or new line

/*   See if have reached the end of a line and normally this 
     would mark the end of a word.  Increment counts and for
     hypen */
     if( inc == '\n' ) {
        if( still_in_word == 0 ) {
            if( num_char > 0 ) {
               num_word++ ; num_word_insent++ ;
               sum_num_char = sum_num_char + num_char ; 
               sum_num_char_sq = sum_num_char_sq + num_char*num_char ; 
          //   Reset number of characters and that we are in gap
               num_char = 0 ;
            }
        }
            still_in_word = 0; 
      }
 } /* End of reading file

/* Now finish up the calcuations */
  mean_char = (float) sum_num_char/num_word ;
  if( num_word > 1 ) {
      rmssq = (sum_num_char_sq - num_word*mean_char*mean_char)/(num_word-1);
      // rms_char = sqrtf(rmssq);
      rms_char = sqrtf((sum_num_char_sq - num_word*mean_char*mean_char)/(num_word-1));
  } else {
      rms_char = 0.0;
  }

  mean_sent = (float) sum_num_word /num_sent ;
  if( num_sent > 1 ) {
     rmssq = (sum_num_word_sq - num_sent*mean_sent*mean_sent)/(num_sent-1);
     rms_sent = sqrtf(rmssq);
  } else {
     rms_sent = 0.0;
  } 

/* Output the results */
  printf("\n12.010 HW03_02: In file %s there are:\n",filename);
  printf("Mean characters per word %5.2f with RMS %5.2f in %4d words;\n",mean_char, rms_char, num_word);
  printf("Mean words per sentence  %5.2f with RMS %5.2f in %4d sentences.\n",mean_sent, rms_sent, num_sent);

  inc = fclose(fp);
}