/* -*- Mode: C; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*-
 * isenglish: make a guess as to whether text argument is English text.
 *            Used for filtering spam sent in other charsets when the
 *            charset specifier is omitted or is in the body rather
 *            than the headers (which seems to be very common in spam).
 * Copyright 2001 by Akkana Peck, akkana@shallowsky.com.
 * You are free to use or modify this program under the GPL.
 */

#include <stdio.h>
#include <ctype.h>
#include <stdlib.h>

/* Table of English letter frequencies per 1000 letters, from
 * http://library.thinkquest.org/28005/flashed/thelab/cryptograms/frequency.shtml
 */
int EnglishLetterFrequencies[] = {
    73, 9, 30, 44, 130, 28, 16, 35, 74, 2, 3, 35, 25,   /* A - M */
    78, 74, 27, 3, 77, 63, 93, 27, 13, 16, 5, 19, 1     /* N - Z */
};

static int Debug = 0;

int
IsVowel(char c)
{
    return (c == 'a' || c == 'e' || c == 'i' || c == 'o' || c == 'u');
}

/* Returns a score from 0 to 100 indicating confidence that it's English */
int
GetScore(char* str)
{
    int letters[26] = {0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0};
    int num = 0;
    int punct = 0;
    int alpha = 0;
    int unprint = 0;
    int vowels = 0;
    int total = 0;
    int words = 0;
    int thiswordlength = 0;
    int avwordlength = 0;
    int score;
    int fromStdin = (str == 0);

    while (fromStdin || (str && *str))
    {
        int c = (fromStdin ? getchar() : *str);

        /* First check for word ends */
        if ((isspace(c) || !isprint(c)) && thiswordlength > 0)
        {
            ++words;
            avwordlength += thiswordlength;
            thiswordlength = 0;
        }

        if (c == EOF)
            break;

        if (isdigit(c))
            ++num;
        else if (isalpha(c))
        {
            ++thiswordlength;
            ++alpha;
            c = tolower(c);
            if (IsVowel(c))
                ++vowels;
            ++letters[c - 'a'];
        }
        else if (isprint(c))
            ++punct;
        else
            ++unprint;
        ++total;
        if (!fromStdin)
            ++str;
    }

    /* Compare alphanum chars to punct chars */
    score = (alpha + num) * 100 / total + 10;

    /* Check word lengths */
    if (words > 0) avwordlength /= words;
    /* Unscientific: just root out extreme cases */
    if (avwordlength < 3 && words > 5)
        score /= 2;
    
    if (score > 100) score = 100;

    if (Debug)
    {
        printf("Score = %d\n", score);
        printf("  (%d alpha, %d digit, %d punct, %d unprintable,\n",
               alpha, num, punct, unprint);
        printf("   %d words, av. word length %d)\n",
               words, avwordlength);
    }
    return score;
}

main(int argc, char **argv)
{
    int score = 0;
    int arg = 1;

    while (argc > 1 && argv[1][0] == '-')
    {
        if (argv[arg][1] == 'd' || argv[arg][1] == 'D')
        {
            printf("Debug mode\n");
            Debug = 1;
        }
        ++arg;
        --argc;
    }

    if (argc > 1)
    {
        int i;
        for (i = 1; i < argc; ++i)
            score += GetScore(argv[arg]);

        score /= (argc - 1);
    }
    else
        score = GetScore(0);

    printf("%d\n", score);
    if (score < 55)
        exit(1);
    exit(0);
}