From a30b8be1ede707f417df6afa5736f6fb75124588 Mon Sep 17 00:00:00 2001
From: Paul Duncan <pabs@pablotron.org>
Date: Sun, 3 Feb 2019 15:39:45 -0500
Subject: sort best, make png optional, clean up debug statements

---
 main.c | 76 ++++++++++++++++++++++++++++++++++++++++++++----------------------
 1 file changed, 51 insertions(+), 25 deletions(-)

(limited to 'main.c')

diff --git a/main.c b/main.c
index 4d86b09..73d5655 100644
--- a/main.c
+++ b/main.c
@@ -20,6 +20,11 @@
 #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;
 
@@ -30,31 +35,45 @@ typedef struct {
     size_t num_empty_clusters;
   } rows[MAX_CLUSTERS - 2];
 
-  km_set_t best[MAX_BEST];
+  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, void *),
+  void (*on_best)(const km_set_t * const, const size_t, const float, void *),
   void * const cb_data
 ) {
-  // if the number of best sets is greater than MAX_BEST, then
-  // find_data->best is actually a ring buffer
-  const bool is_ring = find_data->num_best >= MAX_BEST;
-
-  if (!on_best) {
-    return;
-  }
-
-  // walk best sets
-  for (size_t i = 0; i < MAX_BEST; i++) {
-    // calculate set offset
-    const size_t ofs = i + (is_ring ? (find_data->num_best + 1) : 0);
-
-    // emit set
-    on_best(find_data->best + (ofs % MAX_BEST), i, 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);
+    }
   }
 }
 
@@ -65,7 +84,7 @@ load_on_shape(
 ) {
   km_set_t * const set = cb_data;
 
-  fprintf(stderr, "DEBUG: shape = { %zu, %zu }\n", shape->num_floats, shape->num_ints);
+  D("shape: %zu floats, %zu ints", shape->num_floats, shape->num_ints);
 
   // init set
   if (!km_set_init(set, shape, 100)) {
@@ -150,13 +169,13 @@ find_on_best(
   const km_set_t * const cs,
   void *cb_data
 ) {
-  find_t *find_data = 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);
+  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
@@ -247,15 +266,17 @@ 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, color = %u", rank, color);
+  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);
@@ -288,7 +309,7 @@ save_png(
 int main(int argc, char *argv[]) {
   // check command-line
   if (argc < 2) {
-    fprintf(stderr, "Usage: %s <data>\n", argv[0]);
+    fprintf(stderr, "Usage: %s <data_path> <png_path>\n", argv[0]);
     return EXIT_FAILURE;
   }
 
@@ -310,7 +331,7 @@ int main(int argc, char *argv[]) {
     die("km_set_normalize() failed");
   }
 
-  // find best solution
+  // find best solutions
   if (!km_find(&set, &FIND_CBS, &find_data)) {
     die("km_find()");
   }
@@ -318,8 +339,13 @@ int main(int argc, char *argv[]) {
   // print csv
   print_csv(&find_data);
 
-  // save png of normalized data set and best clusters
-  save_png("data.png", &set, &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);
-- 
cgit v1.2.3