home · contact · privacy
TCE: Set up scenario-specific AI.
authorChristian Heller <c.heller@plomlompom.de>
Sat, 12 Mar 2016 20:02:46 +0000 (21:02 +0100)
committerChristian Heller <c.heller@plomlompom.de>
Sat, 12 Mar 2016 20:02:46 +0000 (21:02 +0100)
libplomrogue.c
plugins/server/TheCrawlingEater.py

index 9f872e1d406158f3a77920de025e6bd9309752c9..68d34311d52fe5844c293a63b721d5a6a3a476df 100644 (file)
@@ -566,6 +566,90 @@ extern uint8_t dijkstra_map()
     return 0;
 }
 
+
+/* 7DRL/TCE addition: movement cost map setting. */
+static uint8_t * TCE_move_cost_map = NULL;
+extern uint8_t TCE_set_movement_cost_map(char * mem_map)
+{
+    uint32_t map_size = maplength * maplength;
+    free(TCE_move_cost_map);
+    TCE_move_cost_map = malloc(map_size * sizeof(uint8_t));
+    uint32_t pos = 0;
+    for (; pos < map_size; pos++)
+    {
+        TCE_move_cost_map[pos] = 0;
+    }
+    if (!TCE_move_cost_map)
+    {
+        return 1;
+    }
+    for (pos = 0; pos < map_size; pos++)
+    {
+        switch(mem_map[pos]) {
+            case '0':
+                TCE_move_cost_map[pos] = 1;
+                break;
+            case '1':
+                TCE_move_cost_map[pos] = 2;
+                break;
+            case '2':
+                TCE_move_cost_map[pos] = 4;
+                break;
+            case '3':
+                TCE_move_cost_map[pos] = 3;
+                break;
+            case '4':
+                TCE_move_cost_map[pos] = 6;
+                break;
+        }
+    }
+    return 0;
+}
+
+
+/* 7DRL/TCE addition: Like dijkstra_map(), but with movement costs applied. */
+extern uint8_t TCE_dijkstra_map_with_movement_cost()
+{
+    if (!score_map || !TCE_move_cost_map)
+    {
+        return 1;
+    }
+    uint16_t max_score = UINT16_MAX - 1;
+    uint32_t map_size = maplength * maplength;
+    uint32_t pos;
+    uint16_t i_scans, neighbors[6], min_neighbor;
+    uint8_t scores_still_changing = 1;
+    uint8_t i_dirs;
+    for (i_scans = 0; scores_still_changing; i_scans++)
+    {
+        scores_still_changing = 0;
+        for (pos = 0; pos < map_size; pos++)
+        {
+            uint16_t score = score_map[pos];
+            uint8_t mov_cost = TCE_move_cost_map[pos];
+            if (mov_cost > 0 && score > i_scans)
+            {
+                get_neighbor_scores(pos, max_score, neighbors);
+                min_neighbor = max_score;
+                for (i_dirs = 0; i_dirs < 6; i_dirs++)
+                {
+                    if (min_neighbor > neighbors[i_dirs])
+                    {
+                        min_neighbor = neighbors[i_dirs];
+                    }
+                }
+                if (score_map[pos] > min_neighbor + mov_cost)
+                {
+                    score_map[pos] = min_neighbor + mov_cost;
+                    scores_still_changing = 1;
+                }
+            }
+        }
+    }
+    return 0;
+}
+
+
 extern uint8_t zero_score_map_where_char_on_memdepthmap(char c,
                                                         char * memdepthmap)
 {
index 658d57db264707c5b114d116b52354e43a1f7786..1bd03d22e63b4779c97c5b6ccf72a81df9781418 100644 (file)
@@ -179,8 +179,7 @@ world_db["test_hole"] = test_hole
 
 
 def test_air(t):
-    if (world_db["wetmap"][t["pos"]] - ord("0")) \
-            + (world_db["MAP"][t["pos"]] - ord("0")) > 5:
+    if world_db["terrain_fullness"](t["pos"]) > 5:
         world_db["die"](t, "You SUFFOCATE")
         return False
     return True
@@ -191,6 +190,7 @@ def die(t, message):
     t["T_LIFEPOINTS"] = 0
     if t == world_db["Things"][0]:
         t["fovmap"] = bytearray(b' ' * (world_db["MAP_LENGTH"] ** 2))
+        t["T_MEMMAP"][t["pos"]] = ord("@")
         log(message)
 world_db["die"] = die
 
@@ -259,7 +259,7 @@ def turn_over():
                     update_map_memory(t)
                     if 0 == tid:
                         return
-                    ai(t)
+                    world_db["ai"](t)
                 if t["T_LIFEPOINTS"]:
                     t["T_PROGRESS"] += 1
                     taid = [a for a in world_db["ThingActions"]
@@ -387,6 +387,222 @@ def write_wetmap():
     return write_map(visible_wetmap, world_db["MAP_LENGTH"])
 
 
+def command_ai():
+    if world_db["WORLD_ACTIVE"]:
+        world_db["ai"](world_db["Things"][0])
+        world_db["turn_over"]()
+
+
+def get_dir_to_target(t, target):
+
+    from server.utils import rand, libpr, c_pointer_to_bytearray
+    from server.config.world_data import symbols_passable
+
+    def zero_score_map_where_char_on_memdepthmap(c):
+        map = c_pointer_to_bytearray(t["T_MEMDEPTHMAP"])
+        if libpr.zero_score_map_where_char_on_memdepthmap(c, map):
+            raise RuntimeError("No score map allocated for "
+                               "zero_score_map_where_char_on_memdepthmap().")
+
+    def set_map_score(pos, score):
+        test = libpr.set_map_score(pos, score)
+        if test:
+            raise RuntimeError("No score map allocated for set_map_score().")
+
+    def set_movement_cost_map():
+        memmap = c_pointer_to_bytearray(t["T_MEMMAP"])
+        if libpr.TCE_set_movement_cost_map(memmap):
+            raise RuntimeError("No movement cost map allocated for "
+                               "set_movement_cost_map().")
+
+    def seeing_thing():
+        def exists(gen):
+            try:
+                next(gen)
+            except StopIteration:
+                return False
+            return True
+        mapsize = world_db["MAP_LENGTH"] ** 2
+        if target == "food" and t["T_MEMMAP"]:
+            return exists(pos for pos in range(mapsize)
+                           if ord("2") < t["T_MEMMAP"][pos] < ord("5"))
+        elif target == "fluid_certain" and t["fovmap"]:
+            return exists(pos for pos in range(mapsize)
+                           if t["fovmap"] == ord("v")
+                           if world_db["MAP"][pos] == ord("0")
+                           if world_db["wetmap"][pos] > ord("0"))
+        elif target == "fluid_potential" and t["T_MEMMAP"] and t["fovmap"]:
+            return exists(pos for pos in range(mapsize)
+                           if t["T_MEMMAP"][pos] == ord("0")
+                           if t["fovmap"] != ord("v"))
+        elif target == "space" and t["T_MEMMAP"] and t["fovmap"]:
+            return exists(pos for pos in range(mapsize)
+                          if ord("0") <= t["T_MEMMAP"][pos] <= ord("2")
+                          if (t["fovmap"] != ord("v")
+                              or world_db["terrain_fullness"](pos) < 5))
+        return False
+
+    def init_score_map():
+        test = libpr.init_score_map()
+        set_movement_cost_map()
+        mapsize = world_db["MAP_LENGTH"] ** 2
+        if test:
+            raise RuntimeError("Malloc error in init_score_map().")
+        if target == "food" and t["T_MEMMAP"]:
+            [set_map_score(pos, 0) for pos in range(mapsize)
+             if ord("2") < t["T_MEMMAP"][pos] < ord("5")]
+        elif target == "fluid_certain" and t["fovmap"]:
+            [set_map_score(pos, 0) for pos in range(mapsize)
+             if t["fovmap"] == ord("v")
+             if world_db["MAP"][pos] == ord("0")
+             if world_db["wetmap"][pos] > ord("0")]
+        elif target == "fluid_potential" and t["T_MEMMAP"] and t["fovmap"]:
+            [set_map_score(pos, 0) for pos in range(mapsize)
+             if t["T_MEMMAP"][pos] == ord("0")
+             if t["fovmap"] != ord("v")]
+        elif target == "space" and t["T_MEMMAP"] and t["fovmap"]:
+            [set_map_score(pos, 0) for pos in range(mapsize)
+             if ord("0") <= t["T_MEMMAP"][pos] <= ord("2")
+             if (t["fovmap"] != ord("v")
+                 or world_db["terrain_fullness"](pos) < 5)]
+        elif target == "search":
+            zero_score_map_where_char_on_memdepthmap(mem_depth_c[0])
+
+    def rand_target_dir(neighbors, cmp, dirs):
+        candidates = []
+        n_candidates = 0
+        for i in range(len(dirs)):
+            if cmp == neighbors[i]:
+                candidates.append(dirs[i])
+                n_candidates += 1
+        return candidates[rand.next() % n_candidates] if n_candidates else 0
+
+    def get_neighbor_scores(dirs, eye_pos):
+        scores = []
+        if libpr.ready_neighbor_scores(eye_pos):
+            raise RuntimeError("No score map allocated for " +
+                               "ready_neighbor_scores.()")
+        for i in range(len(dirs)):
+            scores.append(libpr.get_neighbor_score(i))
+        return scores
+
+    def get_dir_from_neighbors():
+        import math
+        dir_to_target = False
+        dirs = "edcxsw"
+        eye_pos = t["pos"]
+        neighbors = get_neighbor_scores(dirs, eye_pos)
+        minmax_start = 65535 - 1
+        minmax_neighbor = minmax_start
+        for i in range(len(dirs)):
+            if minmax_neighbor > neighbors[i]:
+                minmax_neighbor = neighbors[i]
+        if minmax_neighbor != minmax_start:
+            dir_to_target = rand_target_dir(neighbors, minmax_neighbor, dirs)
+        return dir_to_target, minmax_neighbor
+
+    dir_to_target = False
+    mem_depth_c = b' '
+    run_i = 9 + 1 if "search" == target else 1
+    minmax_neighbor = 0
+    while run_i and not dir_to_target and \
+            ("search" == target or seeing_thing()):
+        run_i -= 1
+        init_score_map()
+        mem_depth_c = b'9' if b' ' == mem_depth_c \
+            else bytes([mem_depth_c[0] - 1])
+        if libpr.TCE_dijkstra_map_with_movement_cost():
+            raise RuntimeError("No score map allocated for dijkstra_map().")
+        dir_to_target, minmax_neighbor = get_dir_from_neighbors()
+        libpr.free_score_map()
+        if dir_to_target and str == type(dir_to_target):
+            action = "move"
+            from server.utils import mv_yx_in_dir_legal
+            move_result = mv_yx_in_dir_legal(dir_to_target, t["T_POSY"],
+                                                            t["T_POSX"])
+            if 1 != move_result[0]:
+                return False, 0
+            pos = (move_result[1] * world_db["MAP_LENGTH"]) + move_result[2]
+            if world_db["MAP"][pos] > ord("2"):
+                action = "eat"
+            t["T_COMMAND"] = [taid for taid in world_db["ThingActions"]
+                              if world_db["ThingActions"][taid]["TA_NAME"]
+                              == action][0]
+            t["T_ARGUMENT"] = ord(dir_to_target)
+    return dir_to_target, minmax_neighbor
+world_db["get_dir_to_target"] = get_dir_to_target
+
+
+def terrain_fullness(pos):
+    return (world_db["MAP"][pos] - ord("0")) + \
+        (world_db["wetmap"][pos] - ord("0"))
+world_db["terrain_fullness"] = terrain_fullness
+
+
+def ai(t):
+
+    if t["T_LIFEPOINTS"] == 0:
+        return
+
+    def standing_on_fluid(t):
+        if world_db["MAP"][t["pos"]] == ord("0") and \
+            world_db["wetmap"][t["pos"]] > ord("0"):
+                return True
+        else:
+            return False
+
+    def thing_action_id(name):
+        return [taid for taid in world_db["ThingActions"]
+                if world_db["ThingActions"][taid]
+                ["TA_NAME"] == name][0]
+
+    t["T_COMMAND"] = thing_action_id("wait")
+    needs = {
+        "safe_pee": (world_db["terrain_fullness"](t["pos"]) * t["T_BLADDER"]) / 4,
+        "safe_drop": (world_db["terrain_fullness"](t["pos"]) * t["T_BOWEL"]) / 4,
+        "food": 33 - t["T_STOMACH"],
+        "fluid_certain": 33 - t["T_KIDNEY"],
+        "fluid_potential": 32 - t["T_KIDNEY"],
+        "search": 1,
+    }
+    from operator import itemgetter
+    needs = sorted(needs.items(), key=itemgetter(1,0))
+    needs.reverse()
+    for need in needs:
+        if need[1] > 0:
+            if need[0] in {"fluid_certain", "fluid_potential"}:
+                if standing_on_fluid(t):
+                    t["T_COMMAND"] = thing_action_id("drink")
+                    return
+                elif t["T_BLADDER"] > 0 and \
+                         world_db["MAP"][t["pos"]] == ord("0"):
+                    t["T_COMMAND"] = thing_action_id("pee")
+                    return
+            elif need[0] in {"safe_pee", "safe_drop"}:
+                action_name = need[0][len("safe_"):]
+                if world_db["terrain_fullness"](t["pos"]) < 4:
+                    t["T_COMMAND"] = thing_action_id(action_name)
+                    return
+                else:
+                    test = world_db["get_dir_to_target"](t, "space")
+                    if test[0]:
+                        if (not test[1] < 5) and \
+                                world_db["terrain_fullness"](t["pos"]) < 5:
+                            t["T_COMMAND"] = thing_action_id(action_name)
+                        return
+                    if t["T_STOMACH"] < 32 and \
+                            world_db["get_dir_to_target"](t, "food")[0]:
+                        return
+                continue
+            if world_db["get_dir_to_target"](t, need[0])[0]:
+                return
+            elif t["T_STOMACH"] < 32 and \
+                    need[0] in {"fluid_certain", "fluid_potential"} and \
+                    world_db["get_dir_to_target"](t, "food")[0]:
+                return
+world_db["ai"] = ai
+
+
 from server.config.io import io_db
 io_db["worldstate_write_order"] += [["T_STOMACH", "player_int"]]
 io_db["worldstate_write_order"] += [["T_KIDNEY", "player_int"]]