/*
  Program to keep a fraction of the number of lines detailed
  in the input file list.  Kept lines are written to the output
  file.  Fraction should be >0.0 and <1.0

  Typically the input file would contain a list of file names
  destined to be migrated (or not) between Lustre OST.

  A random number is thrown and lines are kept if the normalized
  random is <= the desired fraction.
*/

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <time.h>

int main(int argc, char *argv[])
{
  char fullLine[1024];
  int skipLines;
  int totalLines;
  int outputLines;

  char *cPnt;
  FILE *f, *o;
  int tooShort;
  double fraction, keepThese;

  time_t tTime, thisTime;

  if( argc != 4 ) {
    printf ("%d is the wrong number of arguments\n", argc-1);
    printf ("   Please supply: input_file output_file fraction\n");
    exit(1);
  }

  // Check the fraction for validity
  keepThese = atof(argv[3]);
  if ( keepThese <= 0.0 || keepThese > 1.0 ) {
    printf (" Abort: keep fraction %f is out of range\n", keepThese);
    exit(1);
  }

  // Initialize random numbers, etc

  time(&tTime);
  thisTime = mktime(localtime(&tTime));
  srand((unsigned int) thisTime);

  skipLines=0;
  totalLines=0;
  outputLines=0;

  f = fopen(argv[1], "r");
  o = fopen(argv[2], "w");

  while(1) {
    cPnt = fgets( fullLine, 512, f );
    if ( cPnt == NULL ) break;
    totalLines++;
    //    pLast = cPnt + strlen(fullLine);

    tooShort = rand();
    fraction = ((10000.0*tooShort)/RAND_MAX)/10000.0;

    if ( fraction > keepThese ) {
      skipLines++;
      continue;
    }

    fputs(fullLine, o);
    outputLines++;
  }
  fclose(f);
  fclose(o);

  printf( "%d lines read\n%d lines output\n%d lines dropped\n",
	  totalLines, outputLines, skipLines);
  exit(0);
}