uct/uct.c

   1 #include <assert.h>
   2 #include <math.h>
   3 #include <stdio.h>
   4 #include <stdlib.h>
   5 #include <string.h>
   6 #include <time.h>
   7
   8 #define DEBUG
   9
  10 #include "debug.h"
  11 #include "board.h"
  12 #include "gtp.h"
  13 #include "chat.h"
  14 #include "move.h"
  15 #include "mq.h"
  16 #include "joseki/base.h"
  17 #include "playout.h"
  18 #include "playout/moggy.h"
  19 #include "playout/light.h"
  20 #include "tactics/util.h"
  21 #include "timeinfo.h"
  22 #include "uct/dynkomi.h"
  23 #include "uct/internal.h"
  24 #include "uct/plugins.h"
  25 #include "uct/prior.h"
  26 #include "uct/search.h"
  27 #include "uct/slave.h"
  28 #include "uct/tree.h"
  29 #include "uct/uct.h"
  30 #include "uct/walk.h"
  31
  32 struct uct_policy *policy_ucb1_init(struct uct *u, char *arg);
  33 struct uct_policy *policy_ucb1amaf_init(struct uct *u, char *arg, struct board *board);
  34 static void uct_pondering_start(struct uct *u, struct board *b0, struct tree *t, enum stone color);
  35
  36 /* Maximal simulation length. */
  37 #define MC_GAMELEN      MAX_GAMELEN
  38
  39
  40 static void
  41 setup_state(struct uct *u, struct board *b, enum stone color)
  42 {
  43         u->t = tree_init(b, color, u->fast_alloc ? u->max_tree_size : 0,
  44                          u->max_pruned_size, u->pruning_threshold, u->local_tree_aging, u->stats_hbits);
  45         if (u->initial_extra_komi)
  46                 u->t->extra_komi = u->initial_extra_komi;
  47         if (u->force_seed)
  48                 fast_srandom(u->force_seed);
  49         if (UDEBUGL(3))
  50                 fprintf(stderr, "Fresh board with random seed %lu\n", fast_getseed());
  51         if (!u->no_tbook && b->moves == 0) {
  52                 if (color == S_BLACK) {
  53                         tree_load(u->t, b);
  54                 } else if (DEBUGL(0)) {
  55                         fprintf(stderr, "Warning: First move appears to be white\n");
  56                 }
  57         }
  58 }
  59
  60 static void
  61 reset_state(struct uct *u)
  62 {
  63         assert(u->t);
  64         tree_done(u->t); u->t = NULL;
  65 }
  66
  67 static void
  68 setup_dynkomi(struct uct *u, struct board *b, enum stone to_play)
  69 {
  70         if (u->t->use_extra_komi && !u->pondering && u->dynkomi->permove)
  71                 u->t->extra_komi = u->dynkomi->permove(u->dynkomi, b, u->t);
  72         else if (!u->t->use_extra_komi)
  73                 u->t->extra_komi = 0;
  74 }
  75
  76 void
  77 uct_prepare_move(struct uct *u, struct board *b, enum stone color)
  78 {
  79         if (u->t) {
  80                 /* Verify that we have sane state. */
  81                 assert(b->es == u);
  82                 assert(u->t && b->moves);
  83                 if (color != stone_other(u->t->root_color)) {
  84                         fprintf(stderr, "Fatal: Non-alternating play detected %d %d\n",
  85                                 color, u->t->root_color);
  86                         exit(1);
  87                 }
  88                 uct_htable_reset(u->t);
  89
  90         } else {
  91                 /* We need fresh state. */
  92                 b->es = u;
  93                 setup_state(u, b, color);
  94         }
  95
  96         u->ownermap.playouts = 0;
  97         memset(u->ownermap.map, 0, board_size2(b) * sizeof(u->ownermap.map[0]));
  98         u->played_own = u->played_all = 0;
  99 }
 100
 101 static void
 102 dead_group_list(struct uct *u, struct board *b, struct move_queue *mq)
 103 {
 104         enum gj_state gs_array[board_size2(b)];
 105         struct group_judgement gj = { .thres = GJ_THRES, .gs = gs_array };
 106         board_ownermap_judge_groups(b, &u->ownermap, &gj);
 107         groups_of_status(b, &gj, GS_DEAD, mq);
 108 }
 109
 110 bool
 111 uct_pass_is_safe(struct uct *u, struct board *b, enum stone color, bool pass_all_alive)
 112 {
 113         /* Make sure enough playouts are simulated to get a reasonable dead group list. */
 114         while (u->ownermap.playouts < GJ_MINGAMES)
 115                 uct_playout(u, b, color, u->t);
 116
 117         struct move_queue mq = { .moves = 0 };
 118         dead_group_list(u, b, &mq);
 119         if (pass_all_alive) {
 120                 for (unsigned int i = 0; i < mq.moves; i++) {
 121                         if (board_at(b, mq.move[i]) == stone_other(color)) {
 122                                 return false; // We need to remove opponent dead groups first.
 123                         }
 124                 }
 125                 mq.moves = 0; // our dead stones are alive when pass_all_alive is true
 126         }
 127         if (u->allow_losing_pass) {
 128                 foreach_point(b) {
 129                         if (board_at(b, c) == S_OFFBOARD)
 130                                 continue;
 131                         if (board_ownermap_judge_point(&u->ownermap, c, GJ_THRES) == PJ_UNKNOWN) {
 132                                 if (UDEBUGL(3))
 133                                         fprintf(stderr, "uct_pass_is_safe fails at %s[%d]\n", coord2sstr(c, b), c);
 134                                 return false; // Unclear point, clarify first.
 135                         }
 136                 } foreach_point_end;
 137                 return true;
 138         }
 139         return pass_is_safe(b, color, &mq);
 140 }
 141
 142 static char *
 143 uct_printhook_ownermap(struct board *board, coord_t c, char *s, char *end)
 144 {
 145         struct uct *u = board->es;
 146         if (!u) {
 147                 strcat(s, ". ");
 148                 return s + 2;
 149         }
 150         const char chr[] = ":XO,"; // dame, black, white, unclear
 151         const char chm[] = ":xo,";
 152         char ch = chr[board_ownermap_judge_point(&u->ownermap, c, GJ_THRES)];
 153         if (ch == ',') { // less precise estimate then?
 154                 ch = chm[board_ownermap_judge_point(&u->ownermap, c, 0.67)];
 155         }
 156         s += snprintf(s, end - s, "%c ", ch);
 157         return s;
 158 }
 159
 160 static float
 161 uct_owner_map(struct engine *e, struct board *b, coord_t c)
 162 {
 163         struct uct *u = b->es;
 164         return board_ownermap_estimate_point(&u->ownermap, c);
 165 }
 166
 167 static char *
 168 uct_notify_play(struct engine *e, struct board *b, struct move *m, char *enginearg)
 169 {
 170         struct uct *u = e->data;
 171         if (!u->t) {
 172                 /* No state, create one - this is probably game beginning
 173                  * and we need to load the opening tbook right now. */
 174                 uct_prepare_move(u, b, m->color);
 175                 assert(u->t);
 176         }
 177
 178         /* Stop pondering, required by tree_promote_at() */
 179         uct_pondering_stop(u);
 180         if (UDEBUGL(2) && u->slave)
 181                 tree_dump(u->t, u->dumpthres);
 182
 183         if (is_resign(m->coord)) {
 184                 /* Reset state. */
 185                 reset_state(u);
 186                 return NULL;
 187         }
 188
 189         /* Promote node of the appropriate move to the tree root. */
 190         assert(u->t->root);
 191         if (u->t->untrustworthy_tree | !tree_promote_at(u->t, b, m->coord)) {
 192                 if (UDEBUGL(3)) {
 193                         if (u->t->untrustworthy_tree)
 194                                 fprintf(stderr, "Not promoting move node in untrustworthy tree.\n");
 195                         else
 196                                 fprintf(stderr, "Warning: Cannot promote move node! Several play commands in row?\n");
 197                 }
 198                 /* Preserve dynamic komi information, though, that is important. */
 199                 u->initial_extra_komi = u->t->extra_komi;
 200                 reset_state(u);
 201                 return NULL;
 202         }
 203
 204         /* If we are a slave in a distributed engine, start pondering once
 205          * we know which move we actually played. See uct_genmove() about
 206          * the check for pass. */
 207         if (u->pondering_opt && u->slave && m->color == u->my_color && !is_pass(m->coord))
 208                 uct_pondering_start(u, b, u->t, stone_other(m->color));
 209
 210         return NULL;
 211 }
 212
 213 static char *
 214 uct_undo(struct engine *e, struct board *b)
 215 {
 216         struct uct *u = e->data;
 217
 218         if (!u->t) return NULL;
 219         uct_pondering_stop(u);
 220         u->initial_extra_komi = u->t->extra_komi;
 221         reset_state(u);
 222         return NULL;
 223 }
 224
 225 static char *
 226 uct_result(struct engine *e, struct board *b)
 227 {
 228         struct uct *u = e->data;
 229         static char reply[1024];
 230
 231         if (!u->t)
 232                 return NULL;
 233         enum stone color = u->t->root_color;
 234         struct tree_node *n = u->t->root;
 235         snprintf(reply, 1024, "%s %s %d %.2f %.1f",
 236                  stone2str(color), coord2sstr(node_coord(n), b),
 237                  n->u.playouts, tree_node_get_value(u->t, -1, n->u.value),
 238                  u->t->use_extra_komi ? u->t->extra_komi : 0);
 239         return reply;
 240 }
 241
 242 static char *
 243 uct_chat(struct engine *e, struct board *b, bool opponent, char *from, char *cmd)
 244 {
 245         struct uct *u = e->data;
 246
 247         if (!u->t)
 248                 return generic_chat(b, opponent, from, cmd, S_NONE, pass, 0, 1, u->threads, 0.0, 0.0);
 249
 250         struct tree_node *n = u->t->root;
 251         double winrate = tree_node_get_value(u->t, -1, n->u.value);
 252         double extra_komi = u->t->use_extra_komi && fabs(u->t->extra_komi) >= 0.5 ? u->t->extra_komi : 0;
 253
 254         return generic_chat(b, opponent, from, cmd, u->t->root_color, node_coord(n), n->u.playouts, 1,
 255                             u->threads, winrate, extra_komi);
 256 }
 257
 258 static void
 259 uct_dead_group_list(struct engine *e, struct board *b, struct move_queue *mq)
 260 {
 261         struct uct *u = e->data;
 262
 263         /* This means the game is probably over, no use pondering on. */
 264         uct_pondering_stop(u);
 265
 266         if (u->pass_all_alive)
 267                 return; // no dead groups
 268
 269         bool mock_state = false;
 270
 271         if (!u->t) {
 272                 /* No state, but we cannot just back out - we might
 273                  * have passed earlier, only assuming some stones are
 274                  * dead, and then re-connected, only to lose counting
 275                  * when all stones are assumed alive. */
 276                 uct_prepare_move(u, b, S_BLACK); assert(u->t);
 277                 mock_state = true;
 278         }
 279         /* Make sure the ownermap is well-seeded. */
 280         while (u->ownermap.playouts < GJ_MINGAMES)
 281                 uct_playout(u, b, S_BLACK, u->t);
 282         /* Show the ownermap: */
 283         if (DEBUGL(2))
 284                 board_print_custom(b, stderr, uct_printhook_ownermap);
 285
 286         dead_group_list(u, b, mq);
 287
 288         if (mock_state) {
 289                 /* Clean up the mock state in case we will receive
 290                  * a genmove; we could get a non-alternating-move
 291                  * error from uct_prepare_move() in that case otherwise. */
 292                 reset_state(u);
 293         }
 294 }
 295
 296 static void
 297 playout_policy_done(struct playout_policy *p)
 298 {
 299         if (p->done) p->done(p);
 300         if (p->data) free(p->data);
 301         free(p);
 302 }
 303
 304 static void
 305 uct_stop(struct engine *e)
 306 {
 307         /* This is called on game over notification. However, an undo
 308          * and game resume can follow, so don't panic yet and just
 309          * relax and stop thinking so that we don't waste CPU. */
 310         struct uct *u = e->data;
 311         uct_pondering_stop(u);
 312 }
 313
 314 static void
 315 uct_done(struct engine *e)
 316 {
 317         /* This is called on engine reset, especially when clear_board
 318          * is received and new game should begin. */
 319         free(e->comment);
 320
 321         struct uct *u = e->data;
 322         uct_pondering_stop(u);
 323         if (u->t) reset_state(u);
 324         if (u->dynkomi) u->dynkomi->done(u->dynkomi);
 325         free(u->ownermap.map);
 326
 327         if (u->policy) u->policy->done(u->policy);
 328         if (u->random_policy) u->random_policy->done(u->random_policy);
 329         playout_policy_done(u->playout);
 330         uct_prior_done(u->prior);
 331         joseki_done(u->jdict);
 332         pluginset_done(u->plugins);
 333 }
 334
 335
 336
 337 /* Run time-limited MCTS search on foreground. */
 338 static int
 339 uct_search(struct uct *u, struct board *b, struct time_info *ti, enum stone color, struct tree *t, bool print_progress)
 340 {
 341         struct uct_search_state s;
 342         uct_search_start(u, b, color, t, ti, &s);
 343         if (UDEBUGL(2) && s.base_playouts > 0)
 344                 fprintf(stderr, "<pre-simulated %d games>\n", s.base_playouts);
 345
 346         /* The search tree is ctx->t. This is currently == . It is important
 347          * to reference ctx->t directly since the
 348          * thread manager will swap the tree pointer asynchronously. */
 349
 350         /* Now, just periodically poll the search tree. */
 351         /* Note that in case of TD_GAMES, threads will not wait for
 352          * the uct_search_check_stop() signalization. */
 353         while (1) {
 354                 time_sleep(TREE_BUSYWAIT_INTERVAL);
 355                 /* TREE_BUSYWAIT_INTERVAL should never be less than desired time, or the
 356                  * time control is broken. But if it happens to be less, we still search
 357                  * at least 100ms otherwise the move is completely random. */
 358
 359                 int i = uct_search_games(&s);
 360                 /* Print notifications etc. */
 361                 uct_search_progress(u, b, color, t, ti, &s, i);
 362                 /* Check if we should stop the search. */
 363                 if (uct_search_check_stop(u, b, color, t, ti, &s, i))
 364                         break;
 365         }
 366
 367         struct uct_thread_ctx *ctx = uct_search_stop();
 368         if (UDEBUGL(2)) tree_dump(t, u->dumpthres);
 369         if (UDEBUGL(2))
 370                 fprintf(stderr, "(avg score %f/%d; dynkomi's %f/%d value %f/%d)\n",
 371                         t->avg_score.value, t->avg_score.playouts,
 372                         u->dynkomi->score.value, u->dynkomi->score.playouts,
 373                         u->dynkomi->value.value, u->dynkomi->value.playouts);
 374         if (print_progress)
 375                 uct_progress_status(u, t, color, ctx->games, NULL);
 376
 377         if (u->debug_after.playouts > 0) {
 378                 /* Now, start an additional run of playouts, single threaded. */
 379                 struct time_info debug_ti = {
 380                         .period = TT_MOVE,
 381                         .dim = TD_GAMES,
 382                 };
 383                 debug_ti.len.games = t->root->u.playouts + u->debug_after.playouts;
 384
 385                 board_print_custom(b, stderr, uct_printhook_ownermap);
 386                 fprintf(stderr, "--8<-- UCT debug post-run begin (%d:%d) --8<--\n", u->debug_after.level, u->debug_after.playouts);
 387
 388                 int debug_level_save = debug_level;
 389                 int u_debug_level_save = u->debug_level;
 390                 int p_debug_level_save = u->playout->debug_level;
 391                 debug_level = u->debug_after.level;
 392                 u->debug_level = u->debug_after.level;
 393                 u->playout->debug_level = u->debug_after.level;
 394                 uct_halt = false;
 395
 396                 uct_playouts(u, b, color, t, &debug_ti);
 397                 tree_dump(t, u->dumpthres);
 398
 399                 uct_halt = true;
 400                 debug_level = debug_level_save;
 401                 u->debug_level = u_debug_level_save;
 402                 u->playout->debug_level = p_debug_level_save;
 403
 404                 fprintf(stderr, "--8<-- UCT debug post-run finished --8<--\n");
 405         }
 406
 407         u->played_own += ctx->games;
 408         return ctx->games;
 409 }
 410
 411 /* Start pondering background with @color to play. */
 412 static void
 413 uct_pondering_start(struct uct *u, struct board *b0, struct tree *t, enum stone color)
 414 {
 415         if (UDEBUGL(1))
 416                 fprintf(stderr, "Starting to ponder with color %s\n", stone2str(stone_other(color)));
 417         u->pondering = true;
 418
 419         /* We need a local board copy to ponder upon. */
 420         struct board *b = malloc2(sizeof(*b)); board_copy(b, b0);
 421
 422         /* *b0 did not have the genmove'd move played yet. */
 423         struct move m = { node_coord(t->root), t->root_color };
 424         int res = board_play(b, &m);
 425         assert(res >= 0);
 426         setup_dynkomi(u, b, stone_other(m.color));
 427
 428         /* Start MCTS manager thread "headless". */
 429         static struct uct_search_state s;
 430         uct_search_start(u, b, color, t, NULL, &s);
 431 }
 432
 433 /* uct_search_stop() frontend for the pondering (non-genmove) mode, and
 434  * to stop the background search for a slave in the distributed engine. */
 435 void
 436 uct_pondering_stop(struct uct *u)
 437 {
 438         if (!thread_manager_running)
 439                 return;
 440
 441         /* Stop the thread manager. */
 442         struct uct_thread_ctx *ctx = uct_search_stop();
 443         if (UDEBUGL(1)) {
 444                 if (u->pondering) fprintf(stderr, "(pondering) ");
 445                 uct_progress_status(u, ctx->t, ctx->color, ctx->games, NULL);
 446         }
 447         if (u->pondering) {
 448                 free(ctx->b);
 449                 u->pondering = false;
 450         }
 451 }
 452
 453
 454 void
 455 uct_genmove_setup(struct uct *u, struct board *b, enum stone color)
 456 {
 457         if (b->superko_violation) {
 458                 fprintf(stderr, "!!! WARNING: SUPERKO VIOLATION OCCURED BEFORE THIS MOVE\n");
 459                 fprintf(stderr, "Maybe you play with situational instead of positional superko?\n");
 460                 fprintf(stderr, "I'm going to ignore the violation, but note that I may miss\n");
 461                 fprintf(stderr, "some moves valid under this ruleset because of this.\n");
 462                 b->superko_violation = false;
 463         }
 464
 465         uct_prepare_move(u, b, color);
 466
 467         assert(u->t);
 468         u->my_color = color;
 469
 470         /* How to decide whether to use dynkomi in this game? Since we use
 471          * pondering, it's not simple "who-to-play" matter. Decide based on
 472          * the last genmove issued. */
 473         u->t->use_extra_komi = !!(u->dynkomi_mask & color);
 474         setup_dynkomi(u, b, color);
 475
 476         if (b->rules == RULES_JAPANESE)
 477                 u->territory_scoring = true;
 478
 479         /* Make pessimistic assumption about komi for Japanese rules to
 480          * avoid losing by 0.5 when winning by 0.5 with Chinese rules.
 481          * The rules usually give the same winner if the integer part of komi
 482          * is odd so we adjust the komi only if it is even (for a board of
 483          * odd size). We are not trying  to get an exact evaluation for rare
 484          * cases of seki. For details see http://home.snafu.de/jasiek/parity.html */
 485         if (u->territory_scoring && (((int)floor(b->komi) + board_size(b)) & 1)) {
 486                 b->komi += (color == S_BLACK ? 1.0 : -1.0);
 487                 if (UDEBUGL(0))
 488                         fprintf(stderr, "Setting komi to %.1f assuming Japanese rules\n",
 489                                 b->komi);
 490         }
 491 }
 492
 493 static void
 494 uct_live_gfx_hook(struct engine *e)
 495 {
 496         struct uct *u = e->data;
 497         /* Hack: Override reportfreq to get decent update rates in GoGui */
 498         u->reportfreq = 1000;
 499 }
 500
 501 /* Kindof like uct_genmove() but just find the best candidates */
 502 static void
 503 uct_best_moves(struct engine *e, struct board *b, enum stone color)
 504 {
 505         struct time_info ti = { .period = TT_NULL };
 506         double start_time = time_now();
 507         struct uct *u = e->data;
 508         uct_pondering_stop(u);
 509         if (u->t)
 510                 reset_state(u);
 511         uct_genmove_setup(u, b, color);
 512
 513         /* Start the Monte Carlo Tree Search! */
 514         int base_playouts = u->t->root->u.playouts;
 515         int played_games = uct_search(u, b, &ti, color, u->t, false);
 516
 517         coord_t best_coord;
 518         uct_search_result(u, b, color, u->pass_all_alive, played_games, base_playouts, &best_coord);
 519
 520         if (UDEBUGL(2)) {
 521                 double time = time_now() - start_time + 0.000001; /* avoid divide by zero */
 522                 fprintf(stderr, "genmove in %0.2fs (%d games/s, %d games/s/thread)\n",
 523                         time, (int)(played_games/time), (int)(played_games/time/u->threads));
 524         }
 525
 526         uct_progress_status(u, u->t, color, played_games, &best_coord);
 527         reset_state(u);
 528 }
 529
 530 static coord_t *
 531 uct_genmove(struct engine *e, struct board *b, struct time_info *ti, enum stone color, bool pass_all_alive)
 532 {
 533         double start_time = time_now();
 534         struct uct *u = e->data;
 535         u->pass_all_alive |= pass_all_alive;
 536         uct_pondering_stop(u);
 537         uct_genmove_setup(u, b, color);
 538
 539         /* Start the Monte Carlo Tree Search! */
 540         int base_playouts = u->t->root->u.playouts;
 541         int played_games = uct_search(u, b, ti, color, u->t, false);
 542
 543         coord_t best_coord;
 544         struct tree_node *best;
 545         best = uct_search_result(u, b, color, u->pass_all_alive, played_games, base_playouts, &best_coord);
 546
 547         if (UDEBUGL(2)) {
 548                 double time = time_now() - start_time + 0.000001; /* avoid divide by zero */
 549                 fprintf(stderr, "genmove in %0.2fs (%d games/s, %d games/s/thread)\n",
 550                         time, (int)(played_games/time), (int)(played_games/time/u->threads));
 551         }
 552
 553         uct_progress_status(u, u->t, color, played_games, &best_coord);
 554
 555         if (!best) {
 556                 /* Pass or resign. */
 557                 if (is_pass(best_coord))
 558                         u->initial_extra_komi = u->t->extra_komi;
 559                 reset_state(u);
 560                 return coord_copy(best_coord);
 561         }
 562
 563         if (!u->t->untrustworthy_tree) {
 564                 tree_promote_node(u->t, &best);
 565         } else {
 566                 /* Throw away an untrustworthy tree. */
 567                 /* Preserve dynamic komi information, though, that is important. */
 568                 u->initial_extra_komi = u->t->extra_komi;
 569                 reset_state(u);
 570         }
 571
 572         /* After a pass, pondering is harmful for two reasons:
 573          * (i) We might keep pondering even when the game is over.
 574          * Of course this is the case for opponent resign as well.
 575          * (ii) More importantly, the ownermap will get skewed since
 576          * the UCT will start cutting off any playouts. */
 577         if (u->pondering_opt && u->t && !is_pass(node_coord(best))) {
 578                 uct_pondering_start(u, b, u->t, stone_other(color));
 579         }
 580         return coord_copy(best_coord);
 581 }
 582
 583
 584 bool
 585 uct_gentbook(struct engine *e, struct board *b, struct time_info *ti, enum stone color)
 586 {
 587         struct uct *u = e->data;
 588         if (!u->t) uct_prepare_move(u, b, color);
 589         assert(u->t);
 590
 591         if (ti->dim == TD_GAMES) {
 592                 /* Don't count in games that already went into the tbook. */
 593                 ti->len.games += u->t->root->u.playouts;
 594         }
 595         uct_search(u, b, ti, color, u->t, true);
 596
 597         assert(ti->dim == TD_GAMES);
 598         tree_save(u->t, b, ti->len.games / 100);
 599
 600         return true;
 601 }
 602
 603 void
 604 uct_dumptbook(struct engine *e, struct board *b, enum stone color)
 605 {
 606         struct uct *u = e->data;
 607         struct tree *t = tree_init(b, color, u->fast_alloc ? u->max_tree_size : 0,
 608                          u->max_pruned_size, u->pruning_threshold, u->local_tree_aging, 0);
 609         tree_load(t, b);
 610         tree_dump(t, 0);
 611         tree_done(t);
 612 }
 613
 614
 615 floating_t
 616 uct_evaluate_one(struct engine *e, struct board *b, struct time_info *ti, coord_t c, enum stone color)
 617 {
 618         struct uct *u = e->data;
 619
 620         struct board b2;
 621         board_copy(&b2, b);
 622         struct move m = { c, color };
 623         int res = board_play(&b2, &m);
 624         if (res < 0)
 625                 return NAN;
 626         color = stone_other(color);
 627
 628         if (u->t) reset_state(u);
 629         uct_prepare_move(u, &b2, color);
 630         assert(u->t);
 631
 632         floating_t bestval;
 633         uct_search(u, &b2, ti, color, u->t, true);
 634         struct tree_node *best = u->policy->choose(u->policy, u->t->root, &b2, color, resign);
 635         if (!best) {
 636                 bestval = NAN; // the opponent has no reply!
 637         } else {
 638                 bestval = tree_node_get_value(u->t, 1, best->u.value);
 639         }
 640
 641         reset_state(u); // clean our junk
 642
 643         return isnan(bestval) ? NAN : 1.0f - bestval;
 644 }
 645
 646 void
 647 uct_evaluate(struct engine *e, struct board *b, struct time_info *ti, floating_t *vals, enum stone color)
 648 {
 649         for (int i = 0; i < b->flen; i++) {
 650                 if (is_pass(b->f[i]))
 651                         vals[i] = NAN;
 652                 else
 653                         vals[i] = uct_evaluate_one(e, b, ti, b->f[i], color);
 654         }
 655 }
 656
 657
 658 struct uct *
 659 uct_state_init(char *arg, struct board *b)
 660 {
 661         struct uct *u = calloc2(1, sizeof(struct uct));
 662         bool pat_setup = false;
 663
 664         u->debug_level = debug_level;
 665         u->reportfreq = 10000;
 666         u->gamelen = MC_GAMELEN;
 667         u->resign_threshold = 0.2;
 668         u->sure_win_threshold = 0.95;
 669         u->mercymin = 0;
 670         u->significant_threshold = 50;
 671         u->expand_p = 8;
 672         u->dumpthres = 0.01;
 673         u->playout_amaf = true;
 674         u->amaf_prior = false;
 675         u->max_tree_size = 1408ULL * 1048576;
 676         u->fast_alloc = true;
 677         u->pruning_threshold = 0;
 678
 679         u->threads = 1;
 680         u->thread_model = TM_TREEVL;
 681         u->virtual_loss = 1;
 682
 683         u->pondering_opt = true;
 684
 685         u->fuseki_end = 20; // max time at 361*20% = 72 moves (our 36th move, still 99 to play)
 686         u->yose_start = 40; // (100-40-25)*361/100/2 = 63 moves still to play by us then
 687         u->bestr_ratio = 0.02;
 688         // 2.5 is clearly too much, but seems to compensate well for overly stern time allocations.
 689         // TODO: Further tuning and experiments with better time allocation schemes.
 690         u->best2_ratio = 2.5;
 691         // Higher values of max_maintime_ratio sometimes cause severe time trouble in tournaments
 692         // It might be necessary to reduce it to 1.5 on large board, but more tuning is needed.
 693         u->max_maintime_ratio = 2.0;
 694
 695         u->val_scale = 0; u->val_points = 40;
 696         u->dynkomi_interval = 1000;
 697         u->dynkomi_mask = S_BLACK | S_WHITE;
 698
 699         u->tenuki_d = 4;
 700         u->local_tree_aging = 80;
 701         u->local_tree_depth_decay = 1.5;
 702         u->local_tree_eval = LTE_ROOT;
 703         u->local_tree_neival = true;
 704
 705         u->max_slaves = -1;
 706         u->slave_index = -1;
 707         u->stats_delay = 0.01; // 10 ms
 708         u->shared_levels = 1;
 709
 710         u->plugins = pluginset_init(b);
 711
 712         u->jdict = joseki_load(b->size);
 713
 714         if (arg) {
 715                 char *optspec, *next = arg;
 716                 while (*next) {
 717                         optspec = next;
 718                         next += strcspn(next, ",");
 719                         if (*next) { *next++ = 0; } else { *next = 0; }
 720
 721                         char *optname = optspec;
 722                         char *optval = strchr(optspec, '=');
 723                         if (optval) *optval++ = 0;
 724
 725                         /** Basic options */
 726
 727                         if (!strcasecmp(optname, "debug")) {
 728                                 if (optval)
 729                                         u->debug_level = atoi(optval);
 730                                 else
 731                                         u->debug_level++;
 732                         } else if (!strcasecmp(optname, "reporting") && optval) {
 733                                 /* The format of output for detailed progress
 734                                  * information (such as current best move and
 735                                  * its value, etc.). */
 736                                 if (!strcasecmp(optval, "text")) {
 737                                         /* Plaintext traditional output. */
 738                                         u->reporting = UR_TEXT;
 739                                 } else if (!strcasecmp(optval, "json")) {
 740                                         /* JSON output. Implies debug=0. */
 741                                         u->reporting = UR_JSON;
 742                                         u->debug_level = 0;
 743                                 } else if (!strcasecmp(optval, "jsonbig")) {
 744                                         /* JSON output, but much more detailed.
 745                                          * Implies debug=0. */
 746                                         u->reporting = UR_JSON_BIG;
 747                                         u->debug_level = 0;
 748                                 } else {
 749                                         fprintf(stderr, "UCT: Invalid reporting format %s\n", optval);
 750                                         exit(1);
 751                                 }
 752                         } else if (!strcasecmp(optname, "reportfreq") && optval) {
 753                                 /* The progress information line will be shown
 754                                  * every <reportfreq> simulations. */
 755                                 u->reportfreq = atoi(optval);
 756                         } else if (!strcasecmp(optname, "dumpthres") && optval) {
 757                                 /* When dumping the UCT tree on output, include
 758                                  * nodes with at least this many playouts.
 759                                  * (A fraction of the total # of playouts at the
 760                                  * tree root.) */
 761                                 /* Use 0 to list all nodes with at least one
 762                                  * simulation, and -1 to list _all_ nodes. */
 763                                 u->dumpthres = atof(optval);
 764                         } else if (!strcasecmp(optname, "resign_threshold") && optval) {
 765                                 /* Resign when this ratio of games is lost
 766                                  * after GJ_MINGAMES sample is taken. */
 767                                 u->resign_threshold = atof(optval);
 768                         } else if (!strcasecmp(optname, "sure_win_threshold") && optval) {
 769                                 /* Stop reading when this ratio of games is won
 770                                  * after PLAYOUT_EARLY_BREAK_MIN sample is
 771                                  * taken. (Prevents stupid time losses,
 772                                  * friendly to human opponents.) */
 773                                 u->sure_win_threshold = atof(optval);
 774                         } else if (!strcasecmp(optname, "force_seed") && optval) {
 775                                 /* Set RNG seed at the tree setup. */
 776                                 u->force_seed = atoi(optval);
 777                         } else if (!strcasecmp(optname, "no_tbook")) {
 778                                 /* Disable UCT opening tbook. */
 779                                 u->no_tbook = true;
 780                         } else if (!strcasecmp(optname, "pass_all_alive")) {
 781                                 /* Whether to consider passing only after all
 782                                  * dead groups were removed from the board;
 783                                  * this is like all genmoves are in fact
 784                                  * kgs-genmove_cleanup. */
 785                                 u->pass_all_alive = !optval || atoi(optval);
 786                         } else if (!strcasecmp(optname, "allow_losing_pass")) {
 787                                 /* Whether to consider passing in a clear
 788                                  * but losing situation, to be scored as a loss
 789                                  * for us. */
 790                                 u->allow_losing_pass = !optval || atoi(optval);
 791                         } else if (!strcasecmp(optname, "territory_scoring")) {
 792                                 /* Use territory scoring (default is area scoring).
 793                                  * An explicit kgs-rules command overrides this. */
 794                                 u->territory_scoring = !optval || atoi(optval);
 795                         } else if (!strcasecmp(optname, "stones_only")) {
 796                                 /* Do not count eyes. Nice to teach go to kids.
 797                                  * http://strasbourg.jeudego.org/regle_strasbourgeoise.htm */
 798                                 b->rules = RULES_STONES_ONLY;
 799                                 u->pass_all_alive = true;
 800                         } else if (!strcasecmp(optname, "debug_after")) {
 801                                 /* debug_after=9:1000 will make Pachi think under
 802                                  * the normal conditions, but at the point when
 803                                  * a move is to be chosen, the tree is dumped and
 804                                  * another 1000 simulations are run single-threaded
 805                                  * with debug level 9, allowing inspection of Pachi's
 806                                  * behavior after it has thought a lot. */
 807                                 if (optval) {
 808                                         u->debug_after.level = atoi(optval);
 809                                         char *playouts = strchr(optval, ':');
 810                                         if (playouts)
 811                                                 u->debug_after.playouts = atoi(playouts+1);
 812                                         else
 813                                                 u->debug_after.playouts = 1000;
 814                                 } else {
 815                                         u->debug_after.level = 9;
 816                                         u->debug_after.playouts = 1000;
 817                                 }
 818                         } else if (!strcasecmp(optname, "banner") && optval) {
 819                                 /* Additional banner string. This must come as the
 820                                  * last engine parameter. You can use '+' instead
 821                                  * of ' ' if you are wrestling with kgsGtp. */
 822                                 if (*next) *--next = ',';
 823                                 u->banner = strdup(optval);
 824                                 for (char *b = u->banner; *b; b++) {
 825                                         if (*b == '+') *b = ' ';
 826                                 }
 827                                 break;
 828                         } else if (!strcasecmp(optname, "plugin") && optval) {
 829                                 /* Load an external plugin; filename goes before the colon,
 830                                  * extra arguments after the colon. */
 831                                 char *pluginarg = strchr(optval, ':');
 832                                 if (pluginarg)
 833                                         *pluginarg++ = 0;
 834                                 plugin_load(u->plugins, optval, pluginarg);
 835
 836                         /** UCT behavior and policies */
 837
 838                         } else if ((!strcasecmp(optname, "policy")
 839                                 /* Node selection policy. ucb1amaf is the
 840                                  * default policy implementing RAVE, while
 841                                  * ucb1 is the simple exploration/exploitation
 842                                  * policy. Policies can take further extra
 843                                  * options. */
 844                                     || !strcasecmp(optname, "random_policy")) && optval) {
 845                                 /* A policy to be used randomly with small
 846                                  * chance instead of the default policy. */
 847                                 char *policyarg = strchr(optval, ':');
 848                                 struct uct_policy **p = !strcasecmp(optname, "policy") ? &u->policy : &u->random_policy;
 849                                 if (policyarg)
 850                                         *policyarg++ = 0;
 851                                 if (!strcasecmp(optval, "ucb1")) {
 852                                         *p = policy_ucb1_init(u, policyarg);
 853                                 } else if (!strcasecmp(optval, "ucb1amaf")) {
 854                                         *p = policy_ucb1amaf_init(u, policyarg, b);
 855                                 } else {
 856                                         fprintf(stderr, "UCT: Invalid tree policy %s\n", optval);
 857                                         exit(1);
 858                                 }
 859                         } else if (!strcasecmp(optname, "playout") && optval) {
 860                                 /* Random simulation (playout) policy.
 861                                  * moggy is the default policy with large
 862                                  * amount of domain-specific knowledge and
 863                                  * heuristics. light is a simple uniformly
 864                                  * random move selection policy. */
 865                                 char *playoutarg = strchr(optval, ':');
 866                                 if (playoutarg)
 867                                         *playoutarg++ = 0;
 868                                 if (!strcasecmp(optval, "moggy")) {
 869                                         u->playout = playout_moggy_init(playoutarg, b, u->jdict);
 870                                 } else if (!strcasecmp(optval, "light")) {
 871                                         u->playout = playout_light_init(playoutarg, b);
 872                                 } else {
 873                                         fprintf(stderr, "UCT: Invalid playout policy %s\n", optval);
 874                                         exit(1);
 875                                 }
 876                         } else if (!strcasecmp(optname, "prior") && optval) {
 877                                 /* Node priors policy. When expanding a node,
 878                                  * it will seed node values heuristically
 879                                  * (most importantly, based on playout policy
 880                                  * opinion, but also with regard to other
 881                                  * things). See uct/prior.c for details.
 882                                  * Use prior=eqex=0 to disable priors. */
 883                                 u->prior = uct_prior_init(optval, b, u);
 884                         } else if (!strcasecmp(optname, "mercy") && optval) {
 885                                 /* Minimal difference of black/white captures
 886                                  * to stop playout - "Mercy Rule". Speeds up
 887                                  * hopeless playouts at the expense of some
 888                                  * accuracy. */
 889                                 u->mercymin = atoi(optval);
 890                         } else if (!strcasecmp(optname, "gamelen") && optval) {
 891                                 /* Maximum length of single simulation
 892                                  * in moves. */
 893                                 u->gamelen = atoi(optval);
 894                         } else if (!strcasecmp(optname, "expand_p") && optval) {
 895                                 /* Expand UCT nodes after it has been
 896                                  * visited this many times. */
 897                                 u->expand_p = atoi(optval);
 898                         } else if (!strcasecmp(optname, "random_policy_chance") && optval) {
 899                                 /* If specified (N), with probability 1/N, random_policy policy
 900                                  * descend is used instead of main policy descend; useful
 901                                  * if specified policy (e.g. UCB1AMAF) can make unduly biased
 902                                  * choices sometimes, you can fall back to e.g.
 903                                  * random_policy=UCB1. */
 904                                 u->random_policy_chance = atoi(optval);
 905
 906                         /** General AMAF behavior */
 907                         /* (Only relevant if the policy supports AMAF.
 908                          * More variables can be tuned as policy
 909                          * parameters.) */
 910
 911                         } else if (!strcasecmp(optname, "playout_amaf")) {
 912                                 /* Whether to include random playout moves in
 913                                  * AMAF as well. (Otherwise, only tree moves
 914                                  * are included in AMAF. Of course makes sense
 915                                  * only in connection with an AMAF policy.) */
 916                                 /* with-without: 55.5% (+-4.1) */
 917                                 if (optval && *optval == '0')
 918                                         u->playout_amaf = false;
 919                                 else
 920                                         u->playout_amaf = true;
 921                         } else if (!strcasecmp(optname, "playout_amaf_cutoff") && optval) {
 922                                 /* Keep only first N% of playout stage AMAF
 923                                  * information. */
 924                                 u->playout_amaf_cutoff = atoi(optval);
 925                         } else if (!strcasecmp(optname, "amaf_prior") && optval) {
 926                                 /* In node policy, consider prior values
 927                                  * part of the real result term or part
 928                                  * of the AMAF term? */
 929                                 u->amaf_prior = atoi(optval);
 930
 931                         /** Performance and memory management */
 932
 933                         } else if (!strcasecmp(optname, "threads") && optval) {
 934                                 /* By default, Pachi will run with only single
 935                                  * tree search thread! */
 936                                 u->threads = atoi(optval);
 937                         } else if (!strcasecmp(optname, "thread_model") && optval) {
 938                                 if (!strcasecmp(optval, "tree")) {
 939                                         /* Tree parallelization - all threads
 940                                          * grind on the same tree. */
 941                                         u->thread_model = TM_TREE;
 942                                         u->virtual_loss = 0;
 943                                 } else if (!strcasecmp(optval, "treevl")) {
 944                                         /* Tree parallelization, but also
 945                                          * with virtual losses - this discou-
 946                                          * rages most threads choosing the
 947                                          * same tree branches to read. */
 948                                         u->thread_model = TM_TREEVL;
 949                                 } else {
 950                                         fprintf(stderr, "UCT: Invalid thread model %s\n", optval);
 951                                         exit(1);
 952                                 }
 953                         } else if (!strcasecmp(optname, "virtual_loss") && optval) {
 954                                 /* Number of virtual losses added before evaluating a node. */
 955                                 u->virtual_loss = atoi(optval);
 956                         } else if (!strcasecmp(optname, "pondering")) {
 957                                 /* Keep searching even during opponent's turn. */
 958                                 u->pondering_opt = !optval || atoi(optval);
 959                         } else if (!strcasecmp(optname, "max_tree_size") && optval) {
 960                                 /* Maximum amount of memory [MiB] consumed by the move tree.
 961                                  * For fast_alloc it includes the temp tree used for pruning.
 962                                  * Default is 3072 (3 GiB). */
 963                                 u->max_tree_size = atol(optval) * 1048576;
 964                         } else if (!strcasecmp(optname, "fast_alloc")) {
 965                                 u->fast_alloc = !optval || atoi(optval);
 966                         } else if (!strcasecmp(optname, "pruning_threshold") && optval) {
 967                                 /* Force pruning at beginning of a move if the tree consumes
 968                                  * more than this [MiB]. Default is 10% of max_tree_size.
 969                                  * Increase to reduce pruning time overhead if memory is plentiful.
 970                                  * This option is meaningful only for fast_alloc. */
 971                                 u->pruning_threshold = atol(optval) * 1048576;
 972
 973                         /** Time control */
 974
 975                         } else if (!strcasecmp(optname, "best2_ratio") && optval) {
 976                                 /* If set, prolong simulating while
 977                                  * first_best/second_best playouts ratio
 978                                  * is less than best2_ratio. */
 979                                 u->best2_ratio = atof(optval);
 980                         } else if (!strcasecmp(optname, "bestr_ratio") && optval) {
 981                                 /* If set, prolong simulating while
 982                                  * best,best_best_child values delta
 983                                  * is more than bestr_ratio. */
 984                                 u->bestr_ratio = atof(optval);
 985                         } else if (!strcasecmp(optname, "max_maintime_ratio") && optval) {
 986                                 /* If set and while not in byoyomi, prolong simulating no more than
 987                                  * max_maintime_ratio times the normal desired thinking time. */
 988                                 u->max_maintime_ratio = atof(optval);
 989                         } else if (!strcasecmp(optname, "fuseki_end") && optval) {
 990                                 /* At the very beginning it's not worth thinking
 991                                  * too long because the playout evaluations are
 992                                  * very noisy. So gradually increase the thinking
 993                                  * time up to maximum when fuseki_end percent
 994                                  * of the board has been played.
 995                                  * This only applies if we are not in byoyomi. */
 996                                 u->fuseki_end = atoi(optval);
 997                         } else if (!strcasecmp(optname, "yose_start") && optval) {
 998                                 /* When yose_start percent of the board has been
 999                                  * played, or if we are in byoyomi, stop spending
1000                                  * more time and spread the remaining time
1001                                  * uniformly.
1002                                  * Between fuseki_end and yose_start, we spend
1003                                  * a constant proportion of the remaining time
1004                                  * on each move. (yose_start should actually
1005                                  * be much earlier than when real yose start,
1006                                  * but "yose" is a good short name to convey
1007                                  * the idea.) */
1008                                 u->yose_start = atoi(optval);
1009
1010                         /** Dynamic komi */
1011
1012                         } else if (!strcasecmp(optname, "dynkomi") && optval) {
1013                                 /* Dynamic komi approach; there are multiple
1014                                  * ways to adjust komi dynamically throughout
1015                                  * play. We currently support two: */
1016                                 char *dynkomiarg = strchr(optval, ':');
1017                                 if (dynkomiarg)
1018                                         *dynkomiarg++ = 0;
1019                                 if (!strcasecmp(optval, "none")) {
1020                                         u->dynkomi = uct_dynkomi_init_none(u, dynkomiarg, b);
1021                                 } else if (!strcasecmp(optval, "linear")) {
1022                                         /* You should set dynkomi_mask=1 or a very low
1023                                          * handicap_value for white. */
1024                                         u->dynkomi = uct_dynkomi_init_linear(u, dynkomiarg, b);
1025                                 } else if (!strcasecmp(optval, "adaptive")) {
1026                                         /* There are many more knobs to
1027                                          * crank - see uct/dynkomi.c. */
1028                                         u->dynkomi = uct_dynkomi_init_adaptive(u, dynkomiarg, b);
1029                                 } else {
1030                                         fprintf(stderr, "UCT: Invalid dynkomi mode %s\n", optval);
1031                                         exit(1);
1032                                 }
1033                         } else if (!strcasecmp(optname, "dynkomi_mask") && optval) {
1034                                 /* Bitmask of colors the player must be
1035                                  * for dynkomi be applied; the default dynkomi_mask=3 allows
1036                                  * dynkomi even in games where Pachi is white. */
1037                                 u->dynkomi_mask = atoi(optval);
1038                         } else if (!strcasecmp(optname, "dynkomi_interval") && optval) {
1039                                 /* If non-zero, re-adjust dynamic komi
1040                                  * throughout a single genmove reading,
1041                                  * roughly every N simulations. */
1042                                 /* XXX: Does not work with tree
1043                                  * parallelization. */
1044                                 u->dynkomi_interval = atoi(optval);
1045                         } else if (!strcasecmp(optname, "extra_komi") && optval) {
1046                                 /* Initial dynamic komi settings. This
1047                                  * is useful for the adaptive dynkomi
1048                                  * policy as the value to start with
1049                                  * (this is NOT kept fixed) in case
1050                                  * there is not enough time in the search
1051                                  * to adjust the value properly (e.g. the
1052                                  * game was interrupted). */
1053                                 u->initial_extra_komi = atof(optval);
1054
1055                         /** Node value result scaling */
1056
1057                         } else if (!strcasecmp(optname, "val_scale") && optval) {
1058                                 /* How much of the game result value should be
1059                                  * influenced by win size. Zero means it isn't. */
1060                                 u->val_scale = atof(optval);
1061                         } else if (!strcasecmp(optname, "val_points") && optval) {
1062                                 /* Maximum size of win to be scaled into game
1063                                  * result value. Zero means boardsize^2. */
1064                                 u->val_points = atoi(optval) * 2; // result values are doubled
1065                         } else if (!strcasecmp(optname, "val_extra")) {
1066                                 /* If false, the score coefficient will be simply
1067                                  * added to the value, instead of scaling the result
1068                                  * coefficient because of it. */
1069                                 u->val_extra = !optval || atoi(optval);
1070                         } else if (!strcasecmp(optname, "val_byavg")) {
1071                                 /* If true, the score included in the value will
1072                                  * be relative to average score in the current
1073                                  * search episode inst. of jigo. */
1074                                 u->val_byavg = !optval || atoi(optval);
1075                         } else if (!strcasecmp(optname, "val_bytemp")) {
1076                                 /* If true, the value scaling coefficient
1077                                  * is different based on value extremity
1078                                  * (dist. from 0.5), linear between
1079                                  * val_bytemp_min, val_scale. */
1080                                 u->val_bytemp = !optval || atoi(optval);
1081                         } else if (!strcasecmp(optname, "val_bytemp_min") && optval) {
1082                                 /* Minimum val_scale in case of val_bytemp. */
1083                                 u->val_bytemp_min = atof(optval);
1084
1085                         /** Local trees */
1086                         /* (Purely experimental. Does not work - yet!) */
1087
1088                         } else if (!strcasecmp(optname, "local_tree")) {
1089                                 /* Whether to bias exploration by local tree values. */
1090                                 u->local_tree = !optval || atoi(optval);
1091                         } else if (!strcasecmp(optname, "tenuki_d") && optval) {
1092                                 /* Tenuki distance at which to break the local tree. */
1093                                 u->tenuki_d = atoi(optval);
1094                                 if (u->tenuki_d > TREE_NODE_D_MAX + 1) {
1095                                         fprintf(stderr, "uct: tenuki_d must not be larger than TREE_NODE_D_MAX+1 %d\n", TREE_NODE_D_MAX + 1);
1096                                         exit(1);
1097                                 }
1098                         } else if (!strcasecmp(optname, "local_tree_aging") && optval) {
1099                                 /* How much to reduce local tree values between moves. */
1100                                 u->local_tree_aging = atof(optval);
1101                         } else if (!strcasecmp(optname, "local_tree_depth_decay") && optval) {
1102                                 /* With value x>0, during the descent the node
1103                                  * contributes 1/x^depth playouts in
1104                                  * the local tree. I.e., with x>1, nodes more
1105                                  * distant from local situation contribute more
1106                                  * than nodes near the root. */
1107                                 u->local_tree_depth_decay = atof(optval);
1108                         } else if (!strcasecmp(optname, "local_tree_allseq")) {
1109                                 /* If disabled, only complete sequences are stored
1110                                  * in the local tree. If this is on, also
1111                                  * subsequences starting at each move are stored. */
1112                                 u->local_tree_allseq = !optval || atoi(optval);
1113                         } else if (!strcasecmp(optname, "local_tree_neival")) {
1114                                 /* If disabled, local node value is not
1115                                  * computed just based on terminal status
1116                                  * of the coordinate, but also its neighbors. */
1117                                 u->local_tree_neival = !optval || atoi(optval);
1118                         } else if (!strcasecmp(optname, "local_tree_eval")) {
1119                                 /* How is the value inserted in the local tree
1120                                  * determined. */
1121                                 if (!strcasecmp(optval, "root"))
1122                                         /* All moves within a tree branch are
1123                                          * considered wrt. their merit
1124                                          * reaching tachtical goal of making
1125                                          * the first move in the branch
1126                                          * survive. */
1127                                         u->local_tree_eval = LTE_ROOT;
1128                                 else if (!strcasecmp(optval, "each"))
1129                                         /* Each move is considered wrt.
1130                                          * its own survival. */
1131                                         u->local_tree_eval = LTE_EACH;
1132                                 else if (!strcasecmp(optval, "total"))
1133                                         /* The tactical goal is the survival
1134                                          * of all the moves of my color and
1135                                          * non-survival of all the opponent
1136                                          * moves. Local values (and their
1137                                          * inverses) are averaged. */
1138                                         u->local_tree_eval = LTE_TOTAL;
1139                                 else {
1140                                         fprintf(stderr, "uct: unknown local_tree_eval %s\n", optval);
1141                                         exit(1);
1142                                 }
1143                         } else if (!strcasecmp(optname, "local_tree_rootchoose")) {
1144                                 /* If disabled, only moves within the local
1145                                  * tree branch are considered; the values
1146                                  * of the branch roots (i.e. root children)
1147                                  * are ignored. This may make sense together
1148                                  * with eval!=each, we consider only moves
1149                                  * that influence the goal, not the "rating"
1150                                  * of the goal itself. (The real solution
1151                                  * will be probably using criticality to pick
1152                                  * local tree branches.) */
1153                                 u->local_tree_rootchoose = !optval || atoi(optval);
1154
1155                         /** Other heuristics */
1156                         } else if (!strcasecmp(optname, "patterns")) {
1157                                 /* Load pattern database. Various modules
1158                                  * (priors, policies etc.) may make use
1159                                  * of this database. They will request
1160                                  * it automatically in that case, but you
1161                                  * can use this option to tweak the pattern
1162                                  * parameters. */
1163                                 patterns_init(&u->pat, optval, false, true);
1164                                 u->want_pat = pat_setup = true;
1165                         } else if (!strcasecmp(optname, "significant_threshold") && optval) {
1166                                 /* Some heuristics (XXX: none in mainline) rely
1167                                  * on the knowledge of the last "significant"
1168                                  * node in the descent. Such a node is
1169                                  * considered reasonably trustworthy to carry
1170                                  * some meaningful information in the values
1171                                  * of the node and its children. */
1172                                 u->significant_threshold = atoi(optval);
1173
1174                         /** Distributed engine slaves setup */
1175
1176                         } else if (!strcasecmp(optname, "slave")) {
1177                                 /* Act as slave for the distributed engine. */
1178                                 u->slave = !optval || atoi(optval);
1179                         } else if (!strcasecmp(optname, "slave_index") && optval) {
1180                                 /* Optional index if per-slave behavior is desired.
1181                                  * Must be given as index/max */
1182                                 u->slave_index = atoi(optval);
1183                                 char *p = strchr(optval, '/');
1184                                 if (p) u->max_slaves = atoi(++p);
1185                         } else if (!strcasecmp(optname, "shared_nodes") && optval) {
1186                                 /* Share at most shared_nodes between master and slave at each genmoves.
1187                                  * Must use the same value in master and slaves. */
1188                                 u->shared_nodes = atoi(optval);
1189                         } else if (!strcasecmp(optname, "shared_levels") && optval) {
1190                                 /* Share only nodes of level <= shared_levels. */
1191                                 u->shared_levels = atoi(optval);
1192                         } else if (!strcasecmp(optname, "stats_hbits") && optval) {
1193                                 /* Set hash table size to 2^stats_hbits for the shared stats. */
1194                                 u->stats_hbits = atoi(optval);
1195                         } else if (!strcasecmp(optname, "stats_delay") && optval) {
1196                                 /* How long to wait in slave for initial stats to build up before
1197                                  * replying to the genmoves command (in ms) */
1198                                 u->stats_delay = 0.001 * atof(optval);
1199
1200                         /** Presets */
1201
1202                         } else if (!strcasecmp(optname, "maximize_score")) {
1203                                 /* A combination of settings that will make
1204                                  * Pachi try to maximize his points (instead
1205                                  * of playing slack yose) or minimize his loss
1206                                  * (and proceed to counting even when losing). */
1207                                 /* Please note that this preset might be
1208                                  * somewhat weaker than normal Pachi, and the
1209                                  * score maximization is approximate; point size
1210                                  * of win/loss still should not be used to judge
1211                                  * strength of Pachi or the opponent. */
1212                                 /* See README for some further notes. */
1213                                 if (!optval || atoi(optval)) {
1214                                         /* Allow scoring a lost game. */
1215                                         u->allow_losing_pass = true;
1216                                         /* Make Pachi keep his calm when losing
1217                                          * and/or maintain winning marging. */
1218                                         /* Do not play games that are losing
1219                                          * by too much. */
1220                                         /* XXX: komi_ratchet_age=40000 is necessary
1221                                          * with losing_komi_ratchet, but 40000
1222                                          * is somewhat arbitrary value. */
1223                                         char dynkomi_args[] = "losing_komi_ratchet:komi_ratchet_age=60000:no_komi_at_game_end=0:max_losing_komi=30";
1224                                         u->dynkomi = uct_dynkomi_init_adaptive(u, dynkomi_args, b);
1225                                         /* XXX: Values arbitrary so far. */
1226                                         /* XXX: Also, is bytemp sensible when
1227                                          * combined with dynamic komi?! */
1228                                         u->val_scale = 0.01;
1229                                         u->val_bytemp = true;
1230                                         u->val_bytemp_min = 0.001;
1231                                         u->val_byavg = true;
1232                                 }
1233
1234                         } else {
1235                                 fprintf(stderr, "uct: Invalid engine argument %s or missing value\n", optname);
1236                                 exit(1);
1237                         }
1238                 }
1239         }
1240
1241         if (!u->policy)
1242                 u->policy = policy_ucb1amaf_init(u, NULL, b);
1243
1244         if (!!u->random_policy_chance ^ !!u->random_policy) {
1245                 fprintf(stderr, "uct: Only one of random_policy and random_policy_chance is set\n");
1246                 exit(1);
1247         }
1248
1249         if (!u->local_tree) {
1250                 /* No ltree aging. */
1251                 u->local_tree_aging = 1.0f;
1252         }
1253
1254         if (u->fast_alloc) {
1255                 if (u->pruning_threshold < u->max_tree_size / 10)
1256                         u->pruning_threshold = u->max_tree_size / 10;
1257                 if (u->pruning_threshold > u->max_tree_size / 2)
1258                         u->pruning_threshold = u->max_tree_size / 2;
1259
1260                 /* Limit pruning temp space to 20% of memory. Beyond this we discard
1261                  * the nodes and recompute them at the next move if necessary. */
1262                 u->max_pruned_size = u->max_tree_size / 5;
1263                 u->max_tree_size -= u->max_pruned_size;
1264         } else {
1265                 /* Reserve 5% memory in case the background free() are slower
1266                  * than the concurrent allocations. */
1267                 u->max_tree_size -= u->max_tree_size / 20;
1268         }
1269
1270         if (!u->prior)
1271                 u->prior = uct_prior_init(NULL, b, u);
1272
1273         if (!u->playout)
1274                 u->playout = playout_moggy_init(NULL, b, u->jdict);
1275         if (!u->playout->debug_level)
1276                 u->playout->debug_level = u->debug_level;
1277
1278         if (u->want_pat && !pat_setup)
1279                 patterns_init(&u->pat, NULL, false, true);
1280
1281         u->ownermap.map = malloc2(board_size2(b) * sizeof(u->ownermap.map[0]));
1282
1283         if (u->slave) {
1284                 if (!u->stats_hbits) u->stats_hbits = DEFAULT_STATS_HBITS;
1285                 if (!u->shared_nodes) u->shared_nodes = DEFAULT_SHARED_NODES;
1286                 assert(u->shared_levels * board_bits2(b) <= 8 * (int)sizeof(path_t));
1287         }
1288
1289         if (!u->dynkomi)
1290                 u->dynkomi = board_small(b) ? uct_dynkomi_init_none(u, NULL, b)
1291                         : uct_dynkomi_init_linear(u, NULL, b);
1292
1293         /* Some things remain uninitialized for now - the opening tbook
1294          * is not loaded and the tree not set up. */
1295         /* This will be initialized in setup_state() at the first move
1296          * received/requested. This is because right now we are not aware
1297          * about any komi or handicap setup and such. */
1298
1299         return u;
1300 }
1301
1302 struct engine *
1303 engine_uct_init(char *arg, struct board *b)
1304 {
1305         struct uct *u = uct_state_init(arg, b);
1306         struct engine *e = calloc2(1, sizeof(struct engine));
1307         e->name = "UCT";
1308         e->printhook = uct_printhook_ownermap;
1309         e->notify_play = uct_notify_play;
1310         e->chat = uct_chat;
1311         e->undo = uct_undo;
1312         e->result = uct_result;
1313         e->genmove = uct_genmove;
1314         e->genmoves = uct_genmoves;
1315         e->evaluate = uct_evaluate;
1316         e->dead_group_list = uct_dead_group_list;
1317         e->stop = uct_stop;
1318         e->done = uct_done;
1319         e->owner_map = uct_owner_map;
1320         e->best_moves = uct_best_moves;
1321         e->live_gfx_hook = uct_live_gfx_hook;
1322         e->data = u;
1323         if (u->slave)
1324                 e->notify = uct_notify;
1325
1326         const char banner[] = "If you believe you have won but I am still playing, "
1327                 "please help me understand by capturing all dead stones. "
1328                 "Anyone can send me 'winrate' in private chat to get my assessment of the position.";
1329         if (!u->banner) u->banner = "";
1330         e->comment = malloc2(sizeof(banner) + strlen(u->banner) + 1);
1331         sprintf(e->comment, "%s %s", banner, u->banner);
1332
1333         return e;
1334 }