new files

Task/Best-shuffle/C/best-shuffle-1.c

@@ -0,0 +1,111 @@
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include <assert.h>
+#include <limits.h>
+
+#define DEBUG
+
+void best_shuffle(const char* txt, char* result) {
+    const size_t len = strlen(txt);
+    if (len == 0)
+        return;
+
+#ifdef DEBUG
+    // txt and result must have the same length
+    assert(len == strlen(result));
+#endif
+
+    // how many of each character?
+    size_t counts[UCHAR_MAX];
+    memset(counts, '\0', UCHAR_MAX * sizeof(int));
+    size_t fmax = 0;
+    for (size_t i = 0; i < len; i++) {
+        counts[(unsigned char)txt[i]]++;
+        const size_t fnew = counts[(unsigned char)txt[i]];
+        if (fmax < fnew)
+             fmax = fnew;
+    }
+    assert(fmax > 0 && fmax <= len);
+
+    // all character positions, grouped by character
+    size_t *ndx1 = malloc(len * sizeof(size_t));
+    if (ndx1 == NULL)
+        exit(EXIT_FAILURE);
+    for (size_t ch = 0, i = 0; ch < UCHAR_MAX; ch++)
+       if (counts[ch])
+            for (size_t j = 0; j < len; j++)
+                if (ch == (unsigned char)txt[j]) {
+                    ndx1[i] = j;
+                    i++;
+                }
+
+    // regroup them for cycles
+    size_t *ndx2 = malloc(len * sizeof(size_t));
+    if (ndx2 == NULL)
+        exit(EXIT_FAILURE);
+    for (size_t i = 0, n = 0, m = 0; i < len; i++) {
+        ndx2[i] = ndx1[n];
+        n += fmax;
+        if (n >= len) {
+            m++;
+            n = m;
+        }
+    }
+
+    // how long can our cyclic groups be?
+    const size_t grp = 1 + (len - 1) / fmax;
+    assert(grp > 0 && grp <= len);
+
+    // how many of them are full length?
+    const size_t lng = 1 + (len - 1) % fmax;
+    assert(lng > 0 && lng <= len);
+
+    // rotate each group
+    for (size_t i = 0, j = 0; i < fmax; i++) {
+        const size_t first = ndx2[j];
+        const size_t glen = grp - (i < lng ? 0 : 1);
+        for (size_t k = 1; k < glen; k++)
+            ndx1[j + k - 1] = ndx2[j + k];
+        ndx1[j + glen - 1] = first;
+        j += glen;
+    }
+
+    // result is original permuted according to our cyclic groups
+    result[len] = '\0';
+    for (size_t i = 0; i < len; i++)
+        result[ndx2[i]] = txt[ndx1[i]];
+
+    free(ndx1);
+    free(ndx2);
+}
+
+void display(const char* txt1, const char* txt2) {
+    const size_t len = strlen(txt1);
+    assert(len == strlen(txt2));
+    int score = 0;
+    for (size_t i = 0; i < len; i++)
+        if (txt1[i] == txt2[i])
+            score++;
+    (void)printf("%s, %s, (%u)\n", txt1, txt2, score);
+}
+
+int main() {
+    const char* data[] = {"abracadabra", "seesaw", "elk", "grrrrrr",
+                          "up", "a", "aabbbbaa", "", "xxxxx"};
+    const size_t data_len = sizeof(data) / sizeof(data[0]);
+    for (size_t i = 0; i < data_len; i++) {
+        const size_t shuf_len = strlen(data[i]) + 1;
+        char shuf[shuf_len];
+
+#ifdef DEBUG
+        memset(shuf, 0xFF, sizeof shuf);
+        shuf[shuf_len - 1] = '\0';
+#endif
+
+        best_shuffle(data[i], shuf);
+        display(data[i], shuf);
+    }
+
+    return EXIT_SUCCESS;
+}

Task/Best-shuffle/C/best-shuffle-2.c

@@ -0,0 +1,106 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+typedef struct letter_group_t {
+	char c;
+	int count;
+} *letter_p;
+
+struct letter_group_t all_letters[26];
+letter_p letters[26];
+
+/* counts how many of each letter is in a string, used later
+ * to generate permutations
+ */
+int count_letters(const char *s)
+{
+	int i, c;
+	for (i = 0; i < 26; i++) {
+		all_letters[i].count = 0;
+		all_letters[i].c = i + 'a';
+	}
+	while (*s != '\0') {
+		i = *(s++);
+
+		/* don't want to deal with bad inputs */
+		if (i < 'a' || i > 'z') {
+			fprintf(stderr, "Abort: Bad string %s\n", s);
+			exit(1);
+		}
+
+		all_letters[i - 'a'].count++;
+	}
+	for (i = 0, c = 0; i < 26; i++)
+		if (all_letters[i].count)
+			letters[c++] = all_letters + i;
+
+	return c;
+}
+
+int least_overlap, seq_no;
+char out[100], orig[100], best[100];
+
+void permutate(int n_letters, int pos, int overlap)
+{
+	int i, ol;
+	if (pos < 0) {
+                /* if enabled will show all shuffles no worse than current best */
+	//	printf("%s: %d\n", out, overlap);
+
+                /* if better than current best, replace it and reset counter */
+		if (overlap < least_overlap) {
+			least_overlap = overlap;
+			seq_no = 0;
+		}
+
+                /* the Nth best tie has 1/N chance of being kept, so all ties
+                 * have equal chance of being selected even though we don't
+                 * how many there are before hand
+                 */
+		if ( (double)rand() / (RAND_MAX + 1.0) * ++seq_no <= 1)
+			strcpy(best, out);
+
+		return;
+	}
+
+        /* standard "try take the letter; try take not" recursive method */
+	for (i = 0; i < n_letters; i++) {
+		if (!letters[i]->count) continue;
+
+		out[pos] = letters[i]->c;
+		letters[i]->count --;
+		ol = (letters[i]->c == orig[pos]) ? overlap + 1 : overlap;
+
+                /* but don't try options that's already worse than current best */
+		if (ol <= least_overlap)
+			permutate(n_letters, pos - 1, ol);
+
+		letters[i]->count ++;
+	}
+	return;
+}
+
+void do_string(const char *str)
+{
+	least_overlap = strlen(str);
+	strcpy(orig, str);
+
+	seq_no = 0;
+	out[least_overlap] = '\0';
+	least_overlap ++;
+
+	permutate(count_letters(str), least_overlap - 2, 0);
+	printf("%s -> %s, overlap %d\n", str, best, least_overlap);
+}
+
+int main()
+{
+	srand(time(0));
+	do_string("abracadebra");
+	do_string("grrrrrr");
+	do_string("elk");
+	do_string("seesaw");
+	do_string("");
+	return 0;
+}

Task/Best-shuffle/C/best-shuffle-3.c

@@ -0,0 +1,5 @@
+abracadebra -> edbcarabaar, overlap 0
+grrrrrr -> rrgrrrr, overlap 5
+elk -> kel, overlap 0
+seesaw -> ewsesa, overlap 0
+ -> , overlap 0

Task/Best-shuffle/C/best-shuffle-4.c

@@ -0,0 +1,36 @@
+#include <stdio.h>
+#include <string.h>
+
+#define FOR(x, y) for(x = 0; x < y; x++)
+char *best_shuffle(const char *s, int *diff)
+{
+	int i, j = 0, max = 0, l = strlen(s), cnt[128] = {0};
+	char buf[256] = {0}, *r;
+
+	FOR(i, l) if (++cnt[(int)s[i]] > max) max = cnt[(int)s[i]];
+	FOR(i, 128) while (cnt[i]--) buf[j++] = i;
+
+	r = strdup(s);
+	FOR(i, l) FOR(j, l)
+		if (r[i] == buf[j]) {
+			r[i] = buf[(j + max) % l] & ~128;
+			buf[j] |= 128;
+			break;
+		}
+
+	*diff = 0;
+	FOR(i, l) *diff += r[i] == s[i];
+
+	return r;
+}
+
+int main()
+{
+	int i, d;
+	const char *r, *t[] = {"abracadabra", "seesaw", "elk", "grrrrrr", "up", "a", 0};
+	for (i = 0; t[i]; i++) {
+		r = best_shuffle(t[i], &d);
+		printf("%s %s (%d)\n", t[i], r, d);
+	}
+	return 0;
+}