#include <float.h>
#include <math.h>
#include <stdarg.h>
#include <stdio.h>
#include <string.h>

typedef long double real;
typedef unsigned long long count_t;
#define REALRFMT "%Lf"
#define REALPFMT "%.2Lf"
#define REALPFMTA "%*.1Lf"

#if !defined(__cplusplus)
typedef unsigned char bool;
#define false 0
#define true 1
#endif

// Parameters
static count_t file_count = 0;
static bool show_help = false;
static bool force_stats = false;
static bool show_histogram = false;
static bool individual_stats = false;
static bool all_stats = false;
static bool show_version = false;
static char unknown_param = 0;

// Statistics
static count_t count = 0;
static real minimum = DBL_MAX;
static real maximum = -DBL_MAX;
static real value = 0.0;
static real sum = 0.0;
static real square_sum = 0.0;

// Array of data for median, mode and histogram

// Histogram
#define HISTSIZE 10
static size_t histogram[10] = {0};
static real histogram_break_width = 0.0;

void reset_statistics()
{
	count = 0;
	minimum = DBL_MAX;
	maximum = -DBL_MAX;
	value = 0.0;
	sum = 0.0;
	square_sum = 0.0;

	histogram_break_width = 0.0;
	for ( size_t i = 0; i < HISTSIZE; ++i )
		histogram[i] = 0;
}

int eprintf(int result, const char* fmt, ...)
{
	va_list list;
	va_start(list, fmt);
	vfprintf(stderr, fmt, list);
	va_end(list);
	return result;
}

int digits(long long value)
{
	int result = value < 0 ? 2 : 1;
	for ( value /= 10; value; value /= 10 )
		++result;
	return result;
}

int print_help()
{
	printf("%s",
		"Usage: stats [-afhiV] [FILES]\n"
		"Get statistics such as average and standard deviation from\n"
		"numeric data stored in FILES. Options:\n"
		"\n"
		"  -a     print all statistics including median and mode\n"
		"  -f     force to get statistics from no valid files\n"
		"  -h [N] print a histogram with N breaks\n"
		"  -i     print statistics for each individual file\n"
		"  -V     print version of this program\n"
		"\n"
		"Warning: -a and -h options may require high memory consumption\n");

	return 0;
}

int print_version()
{
	printf("%s",
		"stats v2.0 [2013-04-12] Jeisson Hidalgo-Cespedes <jeissonh@gmail.com>\n"
		"\n"
		"stats comes with absolutely no warranty. This is free software, and you are\n"
		"welcome to redistribute it under Creative Commons Attribution 3.0 Unported\n"
		"(CC BY 3.0) license: http://creativecommons.org/licenses/by/3.0/\n");

	return 0;
}

void process_params(int argc, char* argv[])
{
	while ( --argc > 0 )
	{
		char* flags = *++argv;
		if ( *flags == '-' )
		{
			if ( strcmp(flags, "--help") == 0 )
				show_help = true;
			if ( strcmp(flags, "--version") == 0 )
				show_version = true;
			else
				for ( ++flags; *flags; ++flags)
					switch ( *flags )
					{
						case 'a': all_stats = true; break;
						case 'f': force_stats = true; break;
						case 'h': show_histogram = true; break;
						case 'i': individual_stats = true; break;
						case 'V': show_version = true; break;
						default: unknown_param = *flags; break;
					} // switch
		} // if
		else
			++file_count;
	}
}

int process_file(FILE* source, const char* filename, bool build_histogram)
{
	int result = 0;
	while ( true )
	{
		while ( (result = fscanf(source, REALRFMT, &value) ) > 0 )
		{
			if ( build_histogram )
			{
				histogram_break_width = (maximum - minimum) / HISTSIZE;
				size_t index = (value - minimum) / histogram_break_width;
				if ( index >= HISTSIZE ) index = HISTSIZE - 1;
				++histogram[index];
			}
			else
			{
				++count;
				if ( value < minimum ) minimum = value;
				if ( value > maximum ) maximum = value;
				sum += value;
				square_sum += value * value;
			}
		}

		if ( result == EOF )
			return 0;

		if ( force_stats )
			fgetc(source);
		else
			return eprintf(2, "stats:%s: invalid file format\n", filename);
	}
}

void print_histogram()
{
	putchar('\n');
	int maximum_digits = digits(maximum) + 2;
	for ( size_t i = 0; i < HISTSIZE; ++i )
	{
		printf(REALPFMTA " | ", maximum_digits, minimum + i * histogram_break_width);
		for ( size_t j = 0; j < histogram[i]; ++j )
			putchar('*');
		putchar('\n');
	}
}

void print_stats()
{
	if ( count > 0 )
	{
		printf("Count: %llu\n", count);
		printf("Minimum: " REALPFMT "\n", minimum);
		printf("Maximum: " REALPFMT "\n", maximum);
		real average = sum / count;
		printf("Sum: " REALPFMT "\n", sum);
		printf("Average: " REALPFMT "\n", average);
		real variance = (count * average * average * -1 + square_sum) / (count - 1);
		printf("Variance: " REALPFMT "\n", variance);
		printf("Standard deviation: " REALPFMT "\n", sqrtl(variance));
		if ( show_histogram ) print_histogram();
	}
}

int process_filename(const char* filename, bool build_histogram)
{
	if ( *filename == '-' ) return 0;
	FILE* source = fopen(filename, "r");
	if ( ! source ) return eprintf(1, "stats:%s: could not open file\n", filename);
	if ( individual_stats && ! build_histogram ) reset_statistics();
	int result = process_file(source, filename, build_histogram);
	fclose(source);
	if ( individual_stats && ( ! show_histogram || build_histogram ) )
	{
		printf("\n%s:\n", filename);
		print_stats();
	}
	return result;
}

int main(int argc, char* argv[])
{
	int result = 0, temp = 0;
	process_params(argc, argv);

	if ( show_help )
		return print_help();

	if ( show_version )
		return print_version();

	if ( unknown_param )
		return eprintf(1, "stats: -%c unknown option\n", unknown_param);

	if ( show_histogram && file_count == 0 )
		return eprintf(3, "stats: -h option available only for files\n");

	if ( file_count == 0 )
		process_file(stdin, "stdin", false);
	else
		while ( --argc > 0)
			if ( ( temp = process_filename( *++argv, false ) ) != 0 )
				result = temp;
			else if ( show_histogram )
				process_filename(*argv, true);

	if ( ! individual_stats || file_count == 0 ) print_stats();
	return result;
}