path: root/main.c

#include <stdbool.h> // bool
#include <stdio.h> // fprintf()
#include <unistd.h> // EXIT_{FAILURE,SUCCESS}
#include <stdlib.h> // exit()
#include <string.h> // memset()
#include <math.h>  // fabsf()

#define STB_IMAGE_WRITE_IMPLEMENTATION
#define STB_ONLY_PNG
#include "stb_image_write.h"

#include "util.h"
#include "km.h"

#define MAX_CLUSTERS 10
#define NUM_TESTS 100
#define MAX_BEST 4

#define IM_WIDTH 128
#define IM_HEIGHT 128
#define IM_STRIDE (3 * IM_WIDTH)

typedef struct {
  float score;
  km_set_t set;
} best_item_t;

typedef struct {
  km_rand_t rs;

  struct {
    float distance,
          variance,
          cluster_size;
    size_t num_empty_clusters;
  } rows[MAX_CLUSTERS - 2];

  best_item_t best[MAX_BEST];
  size_t num_best;
} find_t;

static int
best_score_cmp(
  const void * const ap,
  const void * const bp
) {
  const best_item_t * const a = ap;
  const best_item_t * const b = bp;

  return (a->score > b->score) ? -1 : 1;
}

static void
find_sort_best(
  find_t * const find_data
) {
  // sort best sets by ascending score (worst to best)
  qsort(
    find_data->best,
    (find_data->num_best % MAX_BEST),
    sizeof(best_item_t),
    best_score_cmp
  );
}

static void
find_each_best(
  const find_t * const find_data,
  void (*on_best)(const km_set_t * const, const size_t, const float, void *),
  void * const cb_data
) {
  if (on_best) {
    // walk best sets and emit each one
    for (size_t i = 0; i < MIN(find_data->num_best, MAX_BEST); i++) {
      on_best(&(find_data->best[i].set), i, find_data->best[i].score, cb_data);
    }
  }
}

static bool
load_on_shape(
  const km_shape_t * const shape,
  void * const cb_data
) {
  km_set_t * const set = cb_data;

  D("shape: %zu floats, %zu ints", shape->num_floats, shape->num_ints);

  // init set
  if (!km_set_init(set, shape, 100)) {
    die("km_set_init() failed");
  }

  // return success
  return true;
}

static bool
load_on_row(
  const float * const floats,
  const int * const ints,
  void * const cb_data
) {
  km_set_t * const set = cb_data;

  // push row
  if (!km_set_push(set, 1, floats, ints)) {
    die("km_set_push_rows() failed");
  }

  // return success
  return true;
}

static void
load_on_error(
  const char * const err,
  void * const cb_data
) {
  UNUSED(cb_data);
  die("load failed: %s", err);
}

static const km_load_cbs_t
LOAD_CBS = {
  .on_shape = load_on_shape,
  .on_row   = load_on_row,
  .on_error = load_on_error,
};

static bool
find_on_init(
  km_set_t * const cs,
  const size_t num_floats,
  const size_t num_clusters,
  void *cb_data
) {
  find_t *data = cb_data;
  return km_set_init_rand_clusters(cs, num_floats, num_clusters, &(data->rs));
}

static bool
find_on_fini(
  km_set_t * const cs,
  void *cb_data
) {
  UNUSED(cb_data);
  km_set_fini(cs);
  return true;
}

static void
find_on_data(
  const km_find_data_t * const data,
  void *cb_data
) {
  find_t * const find_data = cb_data;
  const size_t ofs = data->num_clusters - 2;

  find_data->rows[ofs].distance += data->mean_distance;
  find_data->rows[ofs].variance += data->mean_variance;
  find_data->rows[ofs].cluster_size += data->mean_cluster_size;
  find_data->rows[ofs].num_empty_clusters += data->num_empty_clusters;
}

static bool
find_on_best(
  const float score,
  const km_set_t * const cs,
  void *cb_data
) {
  find_t * const find_data = cb_data;

  D("best score = %0.3f, num_clusters = %zu", score, cs->num_rows);

  // get pointer to destination set
  // (note: data->best is a ring buffer)
  km_set_t *dst = &(find_data->best[find_data->num_best % MAX_BEST].set);

  if (find_data->num_best >= MAX_BEST) {
    // finalize old best data set
    // D("finalizing old best %zu", find_data->num_best);
    km_set_fini(dst);
  }

  // copy data set to best ring buffer
  if (!km_set_copy(dst, cs)) {
    die("km_set_copy()");
  }

  // increment best count
  find_data->num_best++;

  // return success
  return true;
}

// init find config
static const km_find_cbs_t
FIND_CBS = {
  .max_clusters = MAX_CLUSTERS,
  .num_tests    = NUM_TESTS,

  .on_init      = find_on_init,
  .on_fini      = find_on_fini,
  .on_data      = find_on_data,
  .on_best      = find_on_best,
};

static float
get_score(
  const size_t ofs,
  const find_t * const find_data
) {
  // const size_t num_clusters = ofs + 2;
  const float mean_distance = find_data->rows[ofs].distance / NUM_TESTS,
              mean_empty = 1.0 * find_data->rows[ofs].num_empty_clusters / NUM_TESTS;

  return 1.0 / (mean_distance + mean_empty);
}

static void
print_csv_row(
  const size_t i,
  const find_t * const find_data
) {
  const size_t num_clusters = i + 2;
  const float mean_distance = find_data->rows[i].distance / NUM_TESTS,
              mean_variance = find_data->rows[i].variance / NUM_TESTS,
              mean_cluster_size = find_data->rows[i].cluster_size / NUM_TESTS,
              mean_empty_clusters = 1.0 * find_data->rows[i].num_empty_clusters / NUM_TESTS;

  // print result
  printf("%zu,%0.3f,%0.3f,%0.3f,%0.3f,%0.3f\n",
    num_clusters,
    get_score(i, find_data),
    mean_distance,
    mean_variance,
    mean_cluster_size,
    mean_empty_clusters
  );
}

static void
print_csv(
  const find_t * const find_data
) {
  // print headers
  printf(
    "#,"
    "score,"
    "distance,"
    "variance,"
    "cluster_size,"
    "empty_clusters\n"
  );

  for (size_t i = 0; i < MAX_CLUSTERS - 2; i++) {
    print_csv_row(i, find_data);
  }
}

static uint8_t im_data[3 * IM_WIDTH * IM_HEIGHT];

static void
save_on_best(
  const km_set_t * const set,
  const size_t rank,
  const float score,
  void * const cb_data
) {
  UNUSED(cb_data);


  // convert rank to channel brightness
  const uint8_t ch = 0x66 + (0xff - 0x66) * (1.0 * rank) / (MAX_BEST - 1);
  const uint32_t color = (ch & 0xff) << 16;
  // const uint32_t color = 0xff0000;
  D("rank = %zu, score = %0.3f, size = %zu, color = %06x", rank, score, set->num_rows, color);

  // draw clusters
  km_set_draw(set, im_data, IM_WIDTH, IM_HEIGHT, 3, color);
}

static void
save_png(
  const char * const png_path,
  const km_set_t * const set,
  const find_t * const find_data
) {
  // clear image data to white
  memset(im_data, 0xff, sizeof(im_data));

  // draw red points
  km_set_draw(set, im_data, IM_WIDTH, IM_HEIGHT, 1, 0x000000);
  if (!stbi_write_png(png_path, IM_WIDTH, IM_HEIGHT, 3, im_data, IM_STRIDE)) {
    die("stbi_write_png(\"%s\")", png_path);
  }

  // draw best cluster points
  find_each_best(find_data, save_on_best, NULL);

  // save png
  if (!stbi_write_png(png_path, IM_WIDTH, IM_HEIGHT, 3, im_data, IM_STRIDE)) {
    die("stbi_write_png(\"%s\")", png_path);
  }
}

int main(int argc, char *argv[]) {
  // check command-line
  if (argc < 2) {
    fprintf(stderr, "Usage: %s <data_path> <png_path>\n", argv[0]);
    return EXIT_FAILURE;
  }

  // init random seed
  srand(getpid());

  // init find data
  find_t find_data;
  memset(&find_data, 0, sizeof(find_t));
  km_rand_init_system(&(find_data.rs));

  // init data set
  km_set_t set;
  if (!km_load_path(argv[1], &LOAD_CBS, &set)) {
    die("km_load_path() failed");
  }

  if (!km_set_normalize(&set)) {
    die("km_set_normalize() failed");
  }

  // find best solutions
  if (!km_find(&set, &FIND_CBS, &find_data)) {
    die("km_find()");
  }

  // print csv
  print_csv(&find_data);

  // sort best results from lowest to highest
  find_sort_best(&find_data);

  if (argc > 2) {
    // save png of normalized data set and best clusters
    save_png(argv[2], &set, &find_data);
  }

  // finalize data set
  km_set_fini(&set);

  // return success
  return 0;
}