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 "move.h"
  14 #include "mq.h"
  15 #include "playout.h"
  16 #include "playout/elo.h"
  17 #include "playout/moggy.h"
  18 #include "playout/light.h"
  19 #include "tactics.h"
  20 #include "timeinfo.h"
  21 #include "uct/dynkomi.h"
  22 #include "uct/internal.h"
  23 #include "uct/plugins.h"
  24 #include "uct/prior.h"
  25 #include "uct/search.h"
  26 #include "uct/slave.h"
  27 #include "uct/tree.h"
  28 #include "uct/uct.h"
  29 #include "uct/walk.h"
  30
  31 struct uct_policy *policy_ucb1_init(struct uct *u, char *arg);
  32 struct uct_policy *policy_ucb1amaf_init(struct uct *u, char *arg);
  33 static void uct_pondering_start(struct uct *u, struct board *b0, struct tree *t, enum stone color);
  34
  35 /* Maximal simulation length. */
  36 #define MC_GAMELEN      MAX_GAMELEN
  37
  38
  39 static void
  40 setup_state(struct uct *u, struct board *b, enum stone color)
  41 {
  42         u->t = tree_init(b, color, u->fast_alloc ? u->max_tree_size : 0,
  43                          u->local_tree_aging, u->stats_hbits);
  44         if (u->force_seed)
  45                 fast_srandom(u->force_seed);
  46         if (UDEBUGL(0))
  47                 fprintf(stderr, "Fresh board with random seed %lu\n", fast_getseed());
  48         //board_print(b, stderr);
  49         if (!u->no_book && b->moves == 0) {
  50                 assert(color == S_BLACK);
  51                 tree_load(u->t, b);
  52         }
  53 }
  54
  55 static void
  56 reset_state(struct uct *u)
  57 {
  58         assert(u->t);
  59         tree_done(u->t); u->t = NULL;
  60 }
  61
  62 static void
  63 setup_dynkomi(struct uct *u, struct board *b, enum stone to_play)
  64 {
  65         if (u->t->use_extra_komi && !u->pondering && u->dynkomi->permove)
  66                 u->t->extra_komi = u->dynkomi->permove(u->dynkomi, b, u->t);
  67 }
  68
  69 void
  70 uct_prepare_move(struct uct *u, struct board *b, enum stone color)
  71 {
  72         if (u->t) {
  73                 /* Verify that we have sane state. */
  74                 assert(b->es == u);
  75                 assert(u->t && b->moves);
  76                 if (color != stone_other(u->t->root_color)) {
  77                         fprintf(stderr, "Fatal: Non-alternating play detected %d %d\n",
  78                                 color, u->t->root_color);
  79                         exit(1);
  80                 }
  81                 uct_htable_reset(u->t);
  82
  83         } else {
  84                 /* We need fresh state. */
  85                 b->es = u;
  86                 setup_state(u, b, color);
  87         }
  88
  89         u->ownermap.playouts = 0;
  90         memset(u->ownermap.map, 0, board_size2(b) * sizeof(u->ownermap.map[0]));
  91         u->played_own = u->played_all = 0;
  92 }
  93
  94 static void
  95 dead_group_list(struct uct *u, struct board *b, struct move_queue *mq)
  96 {
  97         struct group_judgement gj;
  98         gj.thres = GJ_THRES;
  99         gj.gs = alloca(board_size2(b) * sizeof(gj.gs[0]));
 100         board_ownermap_judge_group(b, &u->ownermap, &gj);
 101         groups_of_status(b, &gj, GS_DEAD, mq);
 102 }
 103
 104 bool
 105 uct_pass_is_safe(struct uct *u, struct board *b, enum stone color, bool pass_all_alive)
 106 {
 107         if (u->ownermap.playouts < GJ_MINGAMES)
 108                 return false;
 109
 110         struct move_queue mq = { .moves = 0 };
 111         dead_group_list(u, b, &mq);
 112         if (pass_all_alive && mq.moves > 0)
 113                 return false; // We need to remove some dead groups first.
 114         return pass_is_safe(b, color, &mq);
 115 }
 116
 117 static char *
 118 uct_printhook_ownermap(struct board *board, coord_t c, char *s, char *end)
 119 {
 120         struct uct *u = board->es;
 121         assert(u);
 122         const char chr[] = ":XO,"; // dame, black, white, unclear
 123         const char chm[] = ":xo,";
 124         char ch = chr[board_ownermap_judge_point(&u->ownermap, c, GJ_THRES)];
 125         if (ch == ',') { // less precise estimate then?
 126                 ch = chm[board_ownermap_judge_point(&u->ownermap, c, 0.67)];
 127         }
 128         s += snprintf(s, end - s, "%c ", ch);
 129         return s;
 130 }
 131
 132 static char *
 133 uct_notify_play(struct engine *e, struct board *b, struct move *m)
 134 {
 135         struct uct *u = e->data;
 136         if (!u->t) {
 137                 /* No state, create one - this is probably game beginning
 138                  * and we need to load the opening book right now. */
 139                 uct_prepare_move(u, b, m->color);
 140                 assert(u->t);
 141         }
 142
 143         /* Stop pondering, required by tree_promote_at() */
 144         uct_pondering_stop(u);
 145         if (UDEBUGL(2) && u->slave)
 146                 tree_dump(u->t, u->dumpthres);
 147
 148         if (is_resign(m->coord)) {
 149                 /* Reset state. */
 150                 reset_state(u);
 151                 return NULL;
 152         }
 153
 154         /* Promote node of the appropriate move to the tree root. */
 155         assert(u->t->root);
 156         if (!tree_promote_at(u->t, b, m->coord)) {
 157                 if (UDEBUGL(0))
 158                         fprintf(stderr, "Warning: Cannot promote move node! Several play commands in row?\n");
 159                 reset_state(u);
 160                 return NULL;
 161         }
 162
 163         /* If we are a slave in a distributed engine, start pondering once
 164          * we know which move we actually played. See uct_genmove() about
 165          * the check for pass. */
 166         if (u->pondering_opt && u->slave && m->color == u->my_color && !is_pass(m->coord))
 167                 uct_pondering_start(u, b, u->t, stone_other(m->color));
 168
 169         return NULL;
 170 }
 171
 172 static char *
 173 uct_chat(struct engine *e, struct board *b, char *cmd)
 174 {
 175         struct uct *u = e->data;
 176         static char reply[1024];
 177
 178         cmd += strspn(cmd, " \n\t");
 179         if (!strncasecmp(cmd, "winrate", 7)) {
 180                 if (!u->t)
 181                         return "no game context (yet?)";
 182                 enum stone color = u->t->root_color;
 183                 struct tree_node *n = u->t->root;
 184                 snprintf(reply, 1024, "In %d playouts at %d threads, %s %s can win with %.2f%% probability",
 185                          n->u.playouts, u->threads, stone2str(color), coord2sstr(n->coord, b),
 186                          tree_node_get_value(u->t, -1, n->u.value) * 100);
 187                 if (u->t->use_extra_komi && abs(u->t->extra_komi) >= 0.5) {
 188                         sprintf(reply + strlen(reply), ", while self-imposing extra komi %.1f",
 189                                 u->t->extra_komi);
 190                 }
 191                 strcat(reply, ".");
 192                 return reply;
 193         }
 194         return NULL;
 195 }
 196
 197 static void
 198 uct_dead_group_list(struct engine *e, struct board *b, struct move_queue *mq)
 199 {
 200         struct uct *u = e->data;
 201
 202         /* This means the game is probably over, no use pondering on. */
 203         uct_pondering_stop(u);
 204
 205         if (u->pass_all_alive)
 206                 return; // no dead groups
 207
 208         bool mock_state = false;
 209
 210         if (!u->t) {
 211                 /* No state, but we cannot just back out - we might
 212                  * have passed earlier, only assuming some stones are
 213                  * dead, and then re-connected, only to lose counting
 214                  * when all stones are assumed alive. */
 215                 uct_prepare_move(u, b, S_BLACK); assert(u->t);
 216                 mock_state = true;
 217         }
 218         /* Make sure the ownermap is well-seeded. */
 219         while (u->ownermap.playouts < GJ_MINGAMES)
 220                 uct_playout(u, b, S_BLACK, u->t);
 221         /* Show the ownermap: */
 222         if (DEBUGL(2))
 223                 board_print_custom(b, stderr, uct_printhook_ownermap);
 224
 225         dead_group_list(u, b, mq);
 226
 227         if (mock_state) {
 228                 /* Clean up the mock state in case we will receive
 229                  * a genmove; we could get a non-alternating-move
 230                  * error from uct_prepare_move() in that case otherwise. */
 231                 reset_state(u);
 232         }
 233 }
 234
 235 static void
 236 playout_policy_done(struct playout_policy *p)
 237 {
 238         if (p->done) p->done(p);
 239         if (p->data) free(p->data);
 240         free(p);
 241 }
 242
 243 static void
 244 uct_done(struct engine *e)
 245 {
 246         /* This is called on engine reset, especially when clear_board
 247          * is received and new game should begin. */
 248         struct uct *u = e->data;
 249         uct_pondering_stop(u);
 250         if (u->t) reset_state(u);
 251         free(u->ownermap.map);
 252
 253         free(u->policy);
 254         free(u->random_policy);
 255         playout_policy_done(u->playout);
 256         uct_prior_done(u->prior);
 257         pluginset_done(u->plugins);
 258 }
 259
 260
 261
 262 /* Run time-limited MCTS search on foreground. */
 263 static int
 264 uct_search(struct uct *u, struct board *b, struct time_info *ti, enum stone color, struct tree *t)
 265 {
 266         struct uct_search_state s;
 267         uct_search_start(u, b, color, t, ti, &s);
 268         if (UDEBUGL(2) && s.base_playouts > 0)
 269                 fprintf(stderr, "<pre-simulated %d games>\n", s.base_playouts);
 270
 271         /* The search tree is ctx->t. This is currently == . It is important
 272          * to reference ctx->t directly since the
 273          * thread manager will swap the tree pointer asynchronously. */
 274
 275         /* Now, just periodically poll the search tree. */
 276         while (1) {
 277                 time_sleep(TREE_BUSYWAIT_INTERVAL);
 278                 /* TREE_BUSYWAIT_INTERVAL should never be less than desired time, or the
 279                  * time control is broken. But if it happens to be less, we still search
 280                  * at least 100ms otherwise the move is completely random. */
 281
 282                 int i = uct_search_games(&s);
 283                 /* Print notifications etc. */
 284                 uct_search_progress(u, b, color, t, ti, &s, i);
 285                 /* Check if we should stop the search. */
 286                 if (uct_search_check_stop(u, b, color, t, ti, &s, i))
 287                         break;
 288         }
 289
 290         struct uct_thread_ctx *ctx = uct_search_stop();
 291         if (UDEBUGL(2)) tree_dump(t, u->dumpthres);
 292         if (UDEBUGL(2))
 293                 fprintf(stderr, "(avg score %f/%d value %f/%d)\n",
 294                         u->dynkomi->score.value, u->dynkomi->score.playouts,
 295                         u->dynkomi->value.value, u->dynkomi->value.playouts);
 296         if (UDEBUGL(0))
 297                 uct_progress_status(u, t, color, ctx->games);
 298
 299         u->played_own += ctx->games;
 300         return ctx->games;
 301 }
 302
 303 /* Start pondering background with @color to play. */
 304 static void
 305 uct_pondering_start(struct uct *u, struct board *b0, struct tree *t, enum stone color)
 306 {
 307         if (UDEBUGL(1))
 308                 fprintf(stderr, "Starting to ponder with color %s\n", stone2str(stone_other(color)));
 309         u->pondering = true;
 310
 311         /* We need a local board copy to ponder upon. */
 312         struct board *b = malloc2(sizeof(*b)); board_copy(b, b0);
 313
 314         /* *b0 did not have the genmove'd move played yet. */
 315         struct move m = { t->root->coord, t->root_color };
 316         int res = board_play(b, &m);
 317         assert(res >= 0);
 318         setup_dynkomi(u, b, stone_other(m.color));
 319
 320         /* Start MCTS manager thread "headless". */
 321         static struct uct_search_state s;
 322         uct_search_start(u, b, color, t, NULL, &s);
 323 }
 324
 325 /* uct_search_stop() frontend for the pondering (non-genmove) mode, and
 326  * to stop the background search for a slave in the distributed engine. */
 327 void
 328 uct_pondering_stop(struct uct *u)
 329 {
 330         if (!thread_manager_running)
 331                 return;
 332
 333         /* Stop the thread manager. */
 334         struct uct_thread_ctx *ctx = uct_search_stop();
 335         if (UDEBUGL(1)) {
 336                 if (u->pondering) fprintf(stderr, "(pondering) ");
 337                 uct_progress_status(u, ctx->t, ctx->color, ctx->games);
 338         }
 339         if (u->pondering) {
 340                 free(ctx->b);
 341                 u->pondering = false;
 342         }
 343 }
 344
 345
 346 void
 347 uct_genmove_setup(struct uct *u, struct board *b, enum stone color)
 348 {
 349         if (b->superko_violation) {
 350                 fprintf(stderr, "!!! WARNING: SUPERKO VIOLATION OCCURED BEFORE THIS MOVE\n");
 351                 fprintf(stderr, "Maybe you play with situational instead of positional superko?\n");
 352                 fprintf(stderr, "I'm going to ignore the violation, but note that I may miss\n");
 353                 fprintf(stderr, "some moves valid under this ruleset because of this.\n");
 354                 b->superko_violation = false;
 355         }
 356
 357         uct_prepare_move(u, b, color);
 358
 359         assert(u->t);
 360         u->my_color = color;
 361
 362         /* How to decide whether to use dynkomi in this game? Since we use
 363          * pondering, it's not simple "who-to-play" matter. Decide based on
 364          * the last genmove issued. */
 365         u->t->use_extra_komi = !!(u->dynkomi_mask & color);
 366         /* Moreover, we do not use extra komi at the game end - we are not
 367          * to fool ourselves at this point. */
 368         if (board_estimated_moves_left(b) <= MIN_MOVES_LEFT)
 369                 u->t->use_extra_komi = false;
 370         setup_dynkomi(u, b, color);
 371
 372         if (b->rules == RULES_JAPANESE)
 373                 u->territory_scoring = true;
 374
 375         /* Make pessimistic assumption about komi for Japanese rules to
 376          * avoid losing by 0.5 when winning by 0.5 with Chinese rules.
 377          * The rules usually give the same winner if the integer part of komi
 378          * is odd so we adjust the komi only if it is even (for a board of
 379          * odd size). We are not trying  to get an exact evaluation for rare
 380          * cases of seki. For details see http://home.snafu.de/jasiek/parity.html */
 381         if (u->territory_scoring && (((int)floor(b->komi) + board_size(b)) & 1)) {
 382                 b->komi += (color == S_BLACK ? 1.0 : -1.0);
 383                 if (UDEBUGL(0))
 384                         fprintf(stderr, "Setting komi to %.1f assuming Japanese rules\n",
 385                                 b->komi);
 386         }
 387 }
 388
 389 static coord_t *
 390 uct_genmove(struct engine *e, struct board *b, struct time_info *ti, enum stone color, bool pass_all_alive)
 391 {
 392         double start_time = time_now();
 393         struct uct *u = e->data;
 394         uct_pondering_stop(u);
 395         uct_genmove_setup(u, b, color);
 396
 397         /* Start the Monte Carlo Tree Search! */
 398         int base_playouts = u->t->root->u.playouts;
 399         int played_games = uct_search(u, b, ti, color, u->t);
 400
 401         coord_t best_coord;
 402         struct tree_node *best;
 403         best = uct_search_result(u, b, color, pass_all_alive, played_games, base_playouts, &best_coord);
 404
 405         if (UDEBUGL(2)) {
 406                 double time = time_now() - start_time + 0.000001; /* avoid divide by zero */
 407                 fprintf(stderr, "genmove in %0.2fs (%d games/s, %d games/s/thread)\n",
 408                         time, (int)(played_games/time), (int)(played_games/time/u->threads));
 409         }
 410
 411         if (!best) {
 412                 /* Pass or resign. */
 413                 reset_state(u);
 414                 return coord_copy(best_coord);
 415         }
 416         tree_promote_node(u->t, &best);
 417
 418         /* After a pass, pondering is harmful for two reasons:
 419          * (i) We might keep pondering even when the game is over.
 420          * Of course this is the case for opponent resign as well.
 421          * (ii) More importantly, the ownermap will get skewed since
 422          * the UCT will start cutting off any playouts. */
 423         if (u->pondering_opt && !is_pass(best->coord)) {
 424                 uct_pondering_start(u, b, u->t, stone_other(color));
 425         }
 426         return coord_copy(best_coord);
 427 }
 428
 429
 430 bool
 431 uct_genbook(struct engine *e, struct board *b, struct time_info *ti, enum stone color)
 432 {
 433         struct uct *u = e->data;
 434         if (!u->t) uct_prepare_move(u, b, color);
 435         assert(u->t);
 436
 437         if (ti->dim == TD_GAMES) {
 438                 /* Don't count in games that already went into the book. */
 439                 ti->len.games += u->t->root->u.playouts;
 440         }
 441         uct_search(u, b, ti, color, u->t);
 442
 443         assert(ti->dim == TD_GAMES);
 444         tree_save(u->t, b, ti->len.games / 100);
 445
 446         return true;
 447 }
 448
 449 void
 450 uct_dumpbook(struct engine *e, struct board *b, enum stone color)
 451 {
 452         struct uct *u = e->data;
 453         struct tree *t = tree_init(b, color, u->fast_alloc ? u->max_tree_size : 0, u->local_tree_aging, 0);
 454         tree_load(t, b);
 455         tree_dump(t, 0);
 456         tree_done(t);
 457 }
 458
 459
 460 struct uct *
 461 uct_state_init(char *arg, struct board *b)
 462 {
 463         struct uct *u = calloc2(1, sizeof(struct uct));
 464         bool using_elo = false;
 465
 466         u->debug_level = debug_level;
 467         u->gamelen = MC_GAMELEN;
 468         u->mercymin = 0;
 469         u->expand_p = 2;
 470         u->dumpthres = 1000;
 471         u->playout_amaf = true;
 472         u->playout_amaf_nakade = false;
 473         u->amaf_prior = false;
 474         u->max_tree_size = 3072ULL * 1048576;
 475
 476         u->dynkomi_mask = S_BLACK;
 477
 478         u->threads = 1;
 479         u->thread_model = TM_TREEVL;
 480         u->virtual_loss = true;
 481
 482         u->fuseki_end = 20; // max time at 361*20% = 72 moves (our 36th move, still 99 to play)
 483         u->yose_start = 40; // (100-40-25)*361/100/2 = 63 moves still to play by us then
 484         u->bestr_ratio = 0.02;
 485         // 2.5 is clearly too much, but seems to compensate well for overly stern time allocations.
 486         // TODO: Further tuning and experiments with better time allocation schemes.
 487         u->best2_ratio = 2.5;
 488
 489         u->val_scale = 0.04; u->val_points = 40;
 490         u->dynkomi_interval = 250;
 491
 492         u->tenuki_d = 4;
 493         u->local_tree_aging = 2;
 494
 495         u->plugins = pluginset_init(b);
 496
 497         if (arg) {
 498                 char *optspec, *next = arg;
 499                 while (*next) {
 500                         optspec = next;
 501                         next += strcspn(next, ",");
 502                         if (*next) { *next++ = 0; } else { *next = 0; }
 503
 504                         char *optname = optspec;
 505                         char *optval = strchr(optspec, '=');
 506                         if (optval) *optval++ = 0;
 507
 508                         if (!strcasecmp(optname, "debug")) {
 509                                 if (optval)
 510                                         u->debug_level = atoi(optval);
 511                                 else
 512                                         u->debug_level++;
 513                         } else if (!strcasecmp(optname, "mercy") && optval) {
 514                                 /* Minimal difference of black/white captures
 515                                  * to stop playout - "Mercy Rule". Speeds up
 516                                  * hopeless playouts at the expense of some
 517                                  * accuracy. */
 518                                 u->mercymin = atoi(optval);
 519                         } else if (!strcasecmp(optname, "gamelen") && optval) {
 520                                 u->gamelen = atoi(optval);
 521                         } else if (!strcasecmp(optname, "expand_p") && optval) {
 522                                 u->expand_p = atoi(optval);
 523                         } else if (!strcasecmp(optname, "dumpthres") && optval) {
 524                                 u->dumpthres = atoi(optval);
 525                         } else if (!strcasecmp(optname, "best2_ratio") && optval) {
 526                                 /* If set, prolong simulating while
 527                                  * first_best/second_best playouts ratio
 528                                  * is less than best2_ratio. */
 529                                 u->best2_ratio = atof(optval);
 530                         } else if (!strcasecmp(optname, "bestr_ratio") && optval) {
 531                                 /* If set, prolong simulating while
 532                                  * best,best_best_child values delta
 533                                  * is more than bestr_ratio. */
 534                                 u->bestr_ratio = atof(optval);
 535                         } else if (!strcasecmp(optname, "playout_amaf")) {
 536                                 /* Whether to include random playout moves in
 537                                  * AMAF as well. (Otherwise, only tree moves
 538                                  * are included in AMAF. Of course makes sense
 539                                  * only in connection with an AMAF policy.) */
 540                                 /* with-without: 55.5% (+-4.1) */
 541                                 if (optval && *optval == '0')
 542                                         u->playout_amaf = false;
 543                                 else
 544                                         u->playout_amaf = true;
 545                         } else if (!strcasecmp(optname, "playout_amaf_nakade")) {
 546                                 /* Whether to include nakade moves from playouts
 547                                  * in the AMAF statistics; this tends to nullify
 548                                  * the playout_amaf effect by adding too much
 549                                  * noise. */
 550                                 if (optval && *optval == '0')
 551                                         u->playout_amaf_nakade = false;
 552                                 else
 553                                         u->playout_amaf_nakade = true;
 554                         } else if (!strcasecmp(optname, "playout_amaf_cutoff") && optval) {
 555                                 /* Keep only first N% of playout stage AMAF
 556                                  * information. */
 557                                 u->playout_amaf_cutoff = atoi(optval);
 558                         } else if ((!strcasecmp(optname, "policy") || !strcasecmp(optname, "random_policy")) && optval) {
 559                                 char *policyarg = strchr(optval, ':');
 560                                 struct uct_policy **p = !strcasecmp(optname, "policy") ? &u->policy : &u->random_policy;
 561                                 if (policyarg)
 562                                         *policyarg++ = 0;
 563                                 if (!strcasecmp(optval, "ucb1")) {
 564                                         *p = policy_ucb1_init(u, policyarg);
 565                                 } else if (!strcasecmp(optval, "ucb1amaf")) {
 566                                         *p = policy_ucb1amaf_init(u, policyarg);
 567                                 } else {
 568                                         fprintf(stderr, "UCT: Invalid tree policy %s\n", optval);
 569                                         exit(1);
 570                                 }
 571                         } else if (!strcasecmp(optname, "playout") && optval) {
 572                                 char *playoutarg = strchr(optval, ':');
 573                                 if (playoutarg)
 574                                         *playoutarg++ = 0;
 575                                 if (!strcasecmp(optval, "moggy")) {
 576                                         u->playout = playout_moggy_init(playoutarg, b);
 577                                 } else if (!strcasecmp(optval, "light")) {
 578                                         u->playout = playout_light_init(playoutarg, b);
 579                                 } else if (!strcasecmp(optval, "elo")) {
 580                                         u->playout = playout_elo_init(playoutarg, b);
 581                                         using_elo = true;
 582                                 } else {
 583                                         fprintf(stderr, "UCT: Invalid playout policy %s\n", optval);
 584                                         exit(1);
 585                                 }
 586                         } else if (!strcasecmp(optname, "prior") && optval) {
 587                                 u->prior = uct_prior_init(optval, b);
 588                         } else if (!strcasecmp(optname, "amaf_prior") && optval) {
 589                                 u->amaf_prior = atoi(optval);
 590                         } else if (!strcasecmp(optname, "threads") && optval) {
 591                                 /* By default, Pachi will run with only single
 592                                  * tree search thread! */
 593                                 u->threads = atoi(optval);
 594                         } else if (!strcasecmp(optname, "thread_model") && optval) {
 595                                 if (!strcasecmp(optval, "tree")) {
 596                                         /* Tree parallelization - all threads
 597                                          * grind on the same tree. */
 598                                         u->thread_model = TM_TREE;
 599                                         u->virtual_loss = false;
 600                                 } else if (!strcasecmp(optval, "treevl")) {
 601                                         /* Tree parallelization, but also
 602                                          * with virtual losses - this discou-
 603                                          * rages most threads choosing the
 604                                          * same tree branches to read. */
 605                                         u->thread_model = TM_TREEVL;
 606                                         u->virtual_loss = true;
 607                                 } else {
 608                                         fprintf(stderr, "UCT: Invalid thread model %s\n", optval);
 609                                         exit(1);
 610                                 }
 611                         } else if (!strcasecmp(optname, "pondering")) {
 612                                 /* Keep searching even during opponent's turn. */
 613                                 u->pondering_opt = !optval || atoi(optval);
 614                         } else if (!strcasecmp(optname, "fuseki_end") && optval) {
 615                                 /* At the very beginning it's not worth thinking
 616                                  * too long because the playout evaluations are
 617                                  * very noisy. So gradually increase the thinking
 618                                  * time up to maximum when fuseki_end percent
 619                                  * of the board has been played.
 620                                  * This only applies if we are not in byoyomi. */
 621                                 u->fuseki_end = atoi(optval);
 622                         } else if (!strcasecmp(optname, "yose_start") && optval) {
 623                                 /* When yose_start percent of the board has been
 624                                  * played, or if we are in byoyomi, stop spending
 625                                  * more time and spread the remaining time
 626                                  * uniformly.
 627                                  * Between fuseki_end and yose_start, we spend
 628                                  * a constant proportion of the remaining time
 629                                  * on each move. (yose_start should actually
 630                                  * be much earlier than when real yose start,
 631                                  * but "yose" is a good short name to convey
 632                                  * the idea.) */
 633                                 u->yose_start = atoi(optval);
 634                         } else if (!strcasecmp(optname, "force_seed") && optval) {
 635                                 u->force_seed = atoi(optval);
 636                         } else if (!strcasecmp(optname, "no_book")) {
 637                                 u->no_book = true;
 638                         } else if (!strcasecmp(optname, "dynkomi") && optval) {
 639                                 /* Dynamic komi approach; there are multiple
 640                                  * ways to adjust komi dynamically throughout
 641                                  * play. We currently support two: */
 642                                 char *dynkomiarg = strchr(optval, ':');
 643                                 if (dynkomiarg)
 644                                         *dynkomiarg++ = 0;
 645                                 if (!strcasecmp(optval, "none")) {
 646                                         u->dynkomi = uct_dynkomi_init_none(u, dynkomiarg, b);
 647                                 } else if (!strcasecmp(optval, "linear")) {
 648                                         u->dynkomi = uct_dynkomi_init_linear(u, dynkomiarg, b);
 649                                 } else if (!strcasecmp(optval, "adaptive")) {
 650                                         u->dynkomi = uct_dynkomi_init_adaptive(u, dynkomiarg, b);
 651                                 } else {
 652                                         fprintf(stderr, "UCT: Invalid dynkomi mode %s\n", optval);
 653                                         exit(1);
 654                                 }
 655                         } else if (!strcasecmp(optname, "dynkomi_mask") && optval) {
 656                                 /* Bitmask of colors the player must be
 657                                  * for dynkomi be applied; you may want
 658                                  * to use dynkomi_mask=3 to allow dynkomi
 659                                  * even in games where Pachi is white. */
 660                                 u->dynkomi_mask = atoi(optval);
 661                         } else if (!strcasecmp(optname, "dynkomi_interval") && optval) {
 662                                 /* If non-zero, re-adjust dynamic komi
 663                                  * throughout a single genmove reading,
 664                                  * roughly every N simulations. */
 665                                 /* XXX: Does not work with tree
 666                                  * parallelization. */
 667                                 u->dynkomi_interval = atoi(optval);
 668                         } else if (!strcasecmp(optname, "val_scale") && optval) {
 669                                 /* How much of the game result value should be
 670                                  * influenced by win size. Zero means it isn't. */
 671                                 u->val_scale = atof(optval);
 672                         } else if (!strcasecmp(optname, "val_points") && optval) {
 673                                 /* Maximum size of win to be scaled into game
 674                                  * result value. Zero means boardsize^2. */
 675                                 u->val_points = atoi(optval) * 2; // result values are doubled
 676                         } else if (!strcasecmp(optname, "val_extra")) {
 677                                 /* If false, the score coefficient will be simply
 678                                  * added to the value, instead of scaling the result
 679                                  * coefficient because of it. */
 680                                 u->val_extra = !optval || atoi(optval);
 681                         } else if (!strcasecmp(optname, "local_tree") && optval) {
 682                                 /* Whether to bias exploration by local tree values
 683                                  * (must be supported by the used policy).
 684                                  * 0: Don't.
 685                                  * 1: Do, value = result.
 686                                  * Try to temper the result:
 687                                  * 2: Do, value = 0.5+(result-expected)/2.
 688                                  * 3: Do, value = 0.5+bzz((result-expected)^2).
 689                                  * 4: Do, value = 0.5+sqrt(result-expected)/2. */
 690                                 u->local_tree = atoi(optval);
 691                         } else if (!strcasecmp(optname, "tenuki_d") && optval) {
 692                                 /* Tenuki distance at which to break the local tree. */
 693                                 u->tenuki_d = atoi(optval);
 694                                 if (u->tenuki_d > TREE_NODE_D_MAX + 1) {
 695                                         fprintf(stderr, "uct: tenuki_d must not be larger than TREE_NODE_D_MAX+1 %d\n", TREE_NODE_D_MAX + 1);
 696                                         exit(1);
 697                                 }
 698                         } else if (!strcasecmp(optname, "local_tree_aging") && optval) {
 699                                 /* How much to reduce local tree values between moves. */
 700                                 u->local_tree_aging = atof(optval);
 701                         } else if (!strcasecmp(optname, "local_tree_allseq")) {
 702                                 /* By default, only complete sequences are stored
 703                                  * in the local tree. If this is on, also
 704                                  * subsequences starting at each move are stored. */
 705                                 u->local_tree_allseq = !optval || atoi(optval);
 706                         } else if (!strcasecmp(optname, "local_tree_playout")) {
 707                                 /* Whether to adjust ELO playout probability
 708                                  * distributions according to matched localtree
 709                                  * information. */
 710                                 u->local_tree_playout = !optval || atoi(optval);
 711                         } else if (!strcasecmp(optname, "local_tree_pseqroot")) {
 712                                 /* By default, when we have no sequence move
 713                                  * to suggest in-playout, we give up. If this
 714                                  * is on, we make probability distribution from
 715                                  * sequences first moves instead. */
 716                                 u->local_tree_pseqroot = !optval || atoi(optval);
 717                         } else if (!strcasecmp(optname, "pass_all_alive")) {
 718                                 /* Whether to consider passing only after all
 719                                  * dead groups were removed from the board;
 720                                  * this is like all genmoves are in fact
 721                                  * kgs-genmove_cleanup. */
 722                                 u->pass_all_alive = !optval || atoi(optval);
 723                         } else if (!strcasecmp(optname, "territory_scoring")) {
 724                                 /* Use territory scoring (default is area scoring).
 725                                  * An explicit kgs-rules command overrides this. */
 726                                 u->territory_scoring = !optval || atoi(optval);
 727                         } else if (!strcasecmp(optname, "random_policy_chance") && optval) {
 728                                 /* If specified (N), with probability 1/N, random_policy policy
 729                                  * descend is used instead of main policy descend; useful
 730                                  * if specified policy (e.g. UCB1AMAF) can make unduly biased
 731                                  * choices sometimes, you can fall back to e.g.
 732                                  * random_policy=UCB1. */
 733                                 u->random_policy_chance = atoi(optval);
 734                         } else if (!strcasecmp(optname, "max_tree_size") && optval) {
 735                                 /* Maximum amount of memory [MiB] consumed by the move tree.
 736                                  * For fast_alloc it includes the temp tree used for pruning.
 737                                  * Default is 3072 (3 GiB). */
 738                                 u->max_tree_size = atol(optval) * 1048576;
 739                         } else if (!strcasecmp(optname, "fast_alloc")) {
 740                                 u->fast_alloc = !optval || atoi(optval);
 741                         } else if (!strcasecmp(optname, "slave")) {
 742                                 /* Act as slave for the distributed engine. */
 743                                 u->slave = !optval || atoi(optval);
 744                         } else if (!strcasecmp(optname, "shared_nodes") && optval) {
 745                                 /* Share at most shared_nodes between master and slave at each genmoves.
 746                                  * Must use the same value in master and slaves. */
 747                                 u->shared_nodes = atoi(optval);
 748                         } else if (!strcasecmp(optname, "shared_levels") && optval) {
 749                                 /* Share only nodes of level <= shared_levels. */
 750                                 u->shared_levels = atoi(optval);
 751                         } else if (!strcasecmp(optname, "stats_hbits") && optval) {
 752                                 /* Set hash table size to 2^stats_hbits for the shared stats. */
 753                                 u->stats_hbits = atoi(optval);
 754                         } else if (!strcasecmp(optname, "banner") && optval) {
 755                                 /* Additional banner string. This must come as the
 756                                  * last engine parameter. */
 757                                 if (*next) *--next = ',';
 758                                 u->banner = strdup(optval);
 759                                 break;
 760                         } else if (!strcasecmp(optname, "plugin") && optval) {
 761                                 /* Load an external plugin; filename goes before the colon,
 762                                  * extra arguments after the colon. */
 763                                 char *pluginarg = strchr(optval, ':');
 764                                 if (pluginarg)
 765                                         *pluginarg++ = 0;
 766                                 plugin_load(u->plugins, optval, pluginarg);
 767                         } else {
 768                                 fprintf(stderr, "uct: Invalid engine argument %s or missing value\n", optname);
 769                                 exit(1);
 770                         }
 771                 }
 772         }
 773
 774         u->resign_ratio = 0.2; /* Resign when most games are lost. */
 775         u->loss_threshold = 0.85; /* Stop reading if after at least 2000 playouts this is best value. */
 776         if (!u->policy)
 777                 u->policy = policy_ucb1amaf_init(u, NULL);
 778
 779         if (!!u->random_policy_chance ^ !!u->random_policy) {
 780                 fprintf(stderr, "uct: Only one of random_policy and random_policy_chance is set\n");
 781                 exit(1);
 782         }
 783
 784         if (!u->local_tree) {
 785                 /* No ltree aging. */
 786                 u->local_tree_aging = 1.0f;
 787         }
 788         if (!using_elo)
 789                 u->local_tree_playout = false;
 790
 791         if (u->fast_alloc)
 792                 u->max_tree_size = (100ULL * u->max_tree_size) / (100 + MIN_FREE_MEM_PERCENT);
 793
 794         if (!u->prior)
 795                 u->prior = uct_prior_init(NULL, b);
 796
 797         if (!u->playout)
 798                 u->playout = playout_moggy_init(NULL, b);
 799         u->playout->debug_level = u->debug_level;
 800
 801         u->ownermap.map = malloc2(board_size2(b) * sizeof(u->ownermap.map[0]));
 802
 803         if (u->slave) {
 804                 if (!u->stats_hbits) u->stats_hbits = DEFAULT_STATS_HBITS;
 805                 if (!u->shared_nodes) u->shared_nodes = DEFAULT_SHARED_NODES;
 806                 if (!u->shared_levels) u->shared_levels = 1;
 807                 assert(u->shared_levels * board_bits2(b) <= 8 * (int)sizeof(path_t));
 808         }
 809
 810         if (!u->dynkomi)
 811                 u->dynkomi = uct_dynkomi_init_linear(u, NULL, b);
 812
 813         /* Some things remain uninitialized for now - the opening book
 814          * is not loaded and the tree not set up. */
 815         /* This will be initialized in setup_state() at the first move
 816          * received/requested. This is because right now we are not aware
 817          * about any komi or handicap setup and such. */
 818
 819         return u;
 820 }
 821
 822 struct engine *
 823 engine_uct_init(char *arg, struct board *b)
 824 {
 825         struct uct *u = uct_state_init(arg, b);
 826         struct engine *e = calloc2(1, sizeof(struct engine));
 827         e->name = "UCT Engine";
 828         e->printhook = uct_printhook_ownermap;
 829         e->notify_play = uct_notify_play;
 830         e->chat = uct_chat;
 831         e->genmove = uct_genmove;
 832         e->genmoves = uct_genmoves;
 833         e->dead_group_list = uct_dead_group_list;
 834         e->done = uct_done;
 835         e->data = u;
 836         if (u->slave)
 837                 e->notify = uct_notify;
 838
 839         const char banner[] = "I'm playing UCT. When I'm losing, I will resign, "
 840                 "if I think I win, I play until you pass. "
 841                 "Anyone can send me 'winrate' in private chat to get my assessment of the position.";
 842         if (!u->banner) u->banner = "";
 843         e->comment = malloc2(sizeof(banner) + strlen(u->banner) + 1);
 844         sprintf(e->comment, "%s %s", banner, u->banner);
 845
 846         return e;
 847 }