bigrams.c

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

static int counts[26][26];

static inline int
lookup(const char s[static 3])
{
    return counts[s[0] - 'a'][s[1] - 'a'];
}

static int
cmp(const void *a, const void *b)
{
    return lookup(b) - lookup(a);
}

static void
compute_table(FILE *in)
{
    char last = ' ';
    char c;
    while ((c = fgetc(in)) != EOF) {
        c = tolower(c);
        if (islower(last) && islower(c))
            counts[last - 'a'][c - 'a']++;
        last = c;
    }
}

static void
print_listing(FILE *out, int cutoff)
{
    char list[26*26][3];
    for (int i = 0; i < 26 * 26; i++) {
        list[i][0] = i / 26 + 'a';
        list[i][1] = i % 26 + 'a';
        list[i][2] = '\0';
    }
    qsort(list, 26 * 26, sizeof(list[0]), cmp);
    for (int i = 0; i < 26 * 26; i++)
        if (lookup(list[i]) >= cutoff)
            fprintf(out, "%s %d\n", list[i], lookup(list[i]));
}

static void
print_table(FILE *out, int cutoff)
{
    fputs("const char table[26][27] = {\n    ", out);
    for (int a = 0; a < 26; a++) {
        fputc('"', out);
        for (int b = 0; b < 26; b++)
            if (counts[a][b] >= cutoff)
                fputc(b + 'a', out);
        fputc('"', out);
        if (a != 25)
            fputs(", ", out);
    }
    fputs("\n};\n", out);
}

int
main(int argc, char **argv)
{
    /* Options */
    enum { LISTING, TABLE } mode = LISTING;
    int cutoff = 1;

    /* Option parsing */
    int option;
    while ((option = getopt(argc, argv, "tlc:")) != -1) {
        switch (option) {
            case 't':
                mode = TABLE;
                break;
            case 'l':
                mode = LISTING;
                break;
            case 'c':
                cutoff = atoi(optarg);
                break;
            default:
                return -1;
        }
    }

    compute_table(stdin);
    switch (mode) {
        case LISTING:
            print_listing(stdout, cutoff);
            break;
        case TABLE:
            print_table(stdout, cutoff);
            break;
    }
    return 0;
}