nace.py

"""
 * The MIT License
 *
 * Copyright (c) 2024 Patrick Hammer
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to deal
 * in the Software without restriction, including without limitation the rights
 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 * copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 * THE SOFTWARE.
 * """

from copy import deepcopy
from collections import deque
from hypothesis import *
from world import *
import sys

#Initializing the memory of the AI
FocusSet = dict([])
rules = set([])
negrules = set([])
worldchange = set([])
RuleEvidence = dict([])
observed_world = [[["." for x in world[BOARD][i]] for i in range(len(world[BOARD]))], world[VALUES], world[TIMES]]
nochange = False

#One observe-learn-plan-action cycle of the AI system
def NACE_Cycle(Time, FocusSet, RuleEvidence, loc, observed_world, rulesin, negrules, oldworld, inject_key=""):
    global nochange
    rulesExcluded = set([])
    rules = deepcopy(rulesin)
    loc, observed_world = World_FieldOfView(Time, loc, observed_world, oldworld)
    Hypothesis_BestSelection(rules, rulesExcluded, RuleEvidence, nochange)
    behavior = ""
    if "manual" not in sys.argv:
        favoured_actions, airis_score, favoured_actions_for_revisit, oldest_age = _Plan(Time, observed_world, rules, actions, customGoal = World_GetObjective())
    else:
        observed_world = [[["." for x in world[BOARD][i]] for i in range(len(world[BOARD]))], world[VALUES], world[TIMES]]
    debuginput = inject_key
    if ("debug" in sys.argv or "manual" in sys.argv) and inject_key == "":
        debuginput = input()
    print("\033[1;1H\033[2J")
    plan = []
    all_actions_tried = False
    if "manual" not in sys.argv:
        actionlist = set(actions)
        #make motorbabbling on curiosity&exploration dependent on whether there are still untried actions
        for rule in rules:
            (precondition, consequence) = rule
            action = precondition[0]
            if action in actionlist:
                actionlist.remove(action)
        all_actions_tried = len(actionlist) == 0
        explore_curiosity_modulator = 1.0 if all_actions_tried else 0.5 #1.0 if all_actions_tried and world_is_novel else 0.5
        curiosity_babble_rate, exploit_babble_rate, explore_babble_rate = (explore_curiosity_modulator, 1.0, explore_curiosity_modulator)
        exploit_babble = random.random() > (exploit_babble_rate if airis_score == float("-inf") else curiosity_babble_rate) #babbling when wanting to achieve something or curious about something, and babbling when exploring:
        explore_babble = random.random() > explore_babble_rate #since it might not know yet about all ops, exploring then can be limited
        if airis_score >= 0.9 or exploit_babble or len(favoured_actions) == 0:
            if not exploit_babble and not explore_babble and oldest_age > 0.0 and airis_score == 1.0 and len(favoured_actions_for_revisit) != 0:
                behavior = "EXPLORE"
                print(behavior, Prettyprint_Plan(favoured_actions_for_revisit), "age:", oldest_age)
                action = favoured_actions_for_revisit[0]
                plan = favoured_actions_for_revisit
            else:
                behavior = "BABBLE"
                actli = deepcopy(actions)
                if drop in actli:
                    actli += [drop, drop] #TODO
                action = random.choice(actli) #motorbabbling
        else:
            behavior = "ACHIEVE" if airis_score == float("-inf") else "CURIOUS"
            print(behavior, Prettyprint_Plan(favoured_actions), end=" "); NACE_PrintScore(airis_score)
            action = favoured_actions[0]
            plan = favoured_actions
    else:
        action = up
    if debuginput == "w":
        action = up
    if debuginput == "s":
        action = down
    if debuginput == "a":
        action = left
    if debuginput == "d":
        action = right
    if debuginput == "b":
        action = pick
    if debuginput == "n":
        action = drop
    if debuginput == "t":
        action = toggle
    if debuginput == 'l':
        for x in rules:
            Prettyprint_rule(RuleEvidence, Hypothesis_TruthValue, x)
        input()
    show_plansteps = debuginput == 'p'
    loc, newworld = World_Move(loc, deepcopy(oldworld), action, plan)
    updateRules = "loadknowledge" not in sys.argv
    rulescopy = deepcopy(rules)
    RuleEvidenceCopy = deepcopy(RuleEvidence)
    if len(plan) > 1:
        updateRules = False
    observed_world_old = deepcopy(observed_world)
    loc, observed_world = World_FieldOfView(Time, loc, observed_world, newworld)
    if observed_world_old[VALUES] == observed_world[VALUES] and observed_world_old[BOARD] == observed_world[BOARD]:
        nochange = True
    else:
        nochange = False
    predicted_world, _, __, values = NACE_Predict(Time, FocusSet, deepcopy(observed_world_old), action, rules)
    if "manual" not in sys.argv:
        print(f"\033[0mWorld t={Time} beliefs={len(rules)}:\033[97;40m")
        World_Print(newworld)
        print("\033[0mMental map:\033[97;44m")
    else:
        print("\033[0mObserved map:\033[97;45m")
    World_Print(observed_world)
    lastplanworld = deepcopy(observed_world)
    planworld = deepcopy(predicted_world)
    if "manual" not in sys.argv:
        print("\033[0mPredicted end:\033[97;41m")
        for i in range(1, len(plan)):
            lastplanworld = deepcopy(planworld)
            planworld, _, __, ___ = NACE_Predict(Time, FocusSet, deepcopy(planworld), plan[i], rules)
            if show_plansteps:
                World_Print(planworld)
        if not show_plansteps:
            World_Print(planworld)
    print("\033[0m")
    if "manual" not in sys.argv:
        FocusSet, RuleEvidence, newrules, newnegrules = _Observe(Time, FocusSet, RuleEvidence, observed_world_old, action, observed_world, rules, negrules, predicted_world)
        if not updateRules:
            newrules = rulescopy
            RuleEvidence = RuleEvidenceCopy
        usedRules = deepcopy(newrules)
        for rule in rulesExcluded: #add again so we won't loose them
            newrules.add(rule)
    else:
        usedRules = newrules = newnegrules = rules
    if debuginput == "u":
        for x in usedRules:
            Prettyprint_rule(RuleEvidence, Hypothesis_TruthValue, x)
        input()
    if debuginput == "x":
        for x in rulesExcluded:
            Prettyprint_rule(RuleEvidence, Hypothesis_TruthValue, x)
        input()
    #print("ALL ACTIONS TRIED?", all_actions_tried)
    return usedRules, FocusSet, RuleEvidence, loc, observed_world, newrules, newnegrules, newworld, debuginput, values, lastplanworld, planworld, behavior, plan

# Apply move to the predicted world model whereby we use the learned tules to decide how grid elements might change most likely
def NACE_Predict(Time, FocusSet, oldworld, action, rules, customGoal = None):
    newworld = deepcopy(oldworld)
    used_rules_sumscore = 0.0
    used_rules_amount = 0
    (positionscores, highesthighscore) = _MatchHypotheses(FocusSet, oldworld, action, rules)
    max_focus = None
    if len(FocusSet) > 0:
        max_focus = max(FocusSet, key=lambda k: FocusSet[k])
    age = 0
    for y in range(height):
        for x in range(width):
            if (y,x) not in positionscores:
                continue
            scores, highscore, rule = positionscores[(y,x)]
            #for rule in rules:
            if _RuleApplicable(scores, highscore, highesthighscore, rule):
                newworld[VALUES] = rule[1][3]
                newworld[BOARD][y][x] = rule[1][2]
                used_rules_sumscore += scores.get(rule, 0.0)
                used_rules_amount += 1
    for y in range(height):
        for x in range(width):
            if (y,x) not in positionscores:
                continue
            if max_focus and newworld[BOARD][y][x] in FocusSet and newworld[BOARD][y][x] == max_focus:
                age = max(age, (Time - newworld[TIMES][y][x]))
    score = used_rules_sumscore/used_rules_amount if used_rules_amount > 0 else 1.0 #AIRIS confidence
    #but if the certaintly predicted world has higher value, then set prediction score to the best it can be
    if (newworld[VALUES][0] == 1 and score == 1.0)  or (customGoal and customGoal(newworld)):
        score = float('-inf')
    #while if the certaintly predicted world has lower value, set prediction score to the worst it can be
    if newworld[VALUES][0] == -1 and score == 1.0:
        score = float('inf')
    return newworld, score, age, newworld[VALUES]

# Plan forward searching for situations of highest reward and if there is no such, then for biggest AIRIS uncertainty (max depth & max queue size obeying breadth first search)
def _Plan(Time, world, rules, actions, max_depth=100, max_queue_len=2000, customGoal = None):
    if "random" in sys.argv: return [random.choice(actions)], float("-inf"), [], 0
    queue = deque([(world, [], 0)])  # Initialize queue with world state, empty action list, and depth 0
    encountered = dict([])
    best_score = float("inf")
    best_actions = []
    best_action_combination_for_revisit = []
    oldest_age = 0.0
    while queue:
        if len(queue) > max_queue_len:
            print("Planning queue bound enforced!")
            break
        current_world, planned_actions, depth = queue.popleft()  # Dequeue from the front
        if depth > max_depth:  # If maximum depth is reached, stop searching
            continue
        world_BOARD_VALUES = World_AsTuple([current_world[BOARD], current_world[VALUES][1:]])
        if world_BOARD_VALUES in encountered and depth >= encountered[world_BOARD_VALUES]:
            continue
        if world_BOARD_VALUES not in encountered or depth < encountered[world_BOARD_VALUES]:
            encountered[world_BOARD_VALUES] = depth
        for action in actions:
            new_world, new_score, new_age, _ = NACE_Predict(Time, FocusSet, deepcopy(current_world), action, rules, customGoal)
            #temporary hack to avoid episode reset issue: if 'x' is close to '.' then put 0.5 (only needed for Minigrid levels)
            if getisMinigridWorld(): #episode reset confuses the system
                for y in range(height-1):
                    for x in range(width-1):
                        if world[BOARD][y][x] == ' ' and new_world[BOARD][y][x] == 'x' and new_world[BOARD][y+1][x] == '.' or \
                           new_world[BOARD][y][x] == '.' and world[BOARD][y+1][x] == ' ' and new_world[BOARD][y+1][x] == 'x' or \
                           world[BOARD][y][x] == ' ' and new_world[BOARD][y][x] == 'x' and new_world[BOARD][y][x+1] == '.' or \
                           new_world[BOARD][y][x] == '.' and world[BOARD][y][x+1] == ' ' and new_world[BOARD][y][x+1] == 'x':
                            new_score = 0.5
            if new_world == current_world or new_score == float("inf"):
                continue
            new_Planned_actions = planned_actions + [action]
            if new_score < best_score or (new_score == best_score and len(new_Planned_actions) < len(best_actions)):
                best_actions = new_Planned_actions
                best_score = new_score
            if new_age > oldest_age or (new_age == oldest_age and len(new_Planned_actions) < len(best_action_combination_for_revisit)):
                best_action_combination_for_revisit = new_Planned_actions
                oldest_age = new_age
            if new_score == 1.0:
                queue.append((new_world, new_Planned_actions, depth + 1))  # Enqueue children at the end
            if new_score == float("-inf"):
                queue = []; break
    return best_actions, best_score, best_action_combination_for_revisit, oldest_age

#Whether the grid cell has been observed now (not all have been, due to partial observability)
def _IsPresentlyObserved(Time, world, y, x):
    return Time - world[TIMES][y][x] == 0

#Build adjacent consideration set
def _AddToAdjacentSet(consideredSets, newEntry, MaxCapacity, CanCreateNewSet):
    (y, x) = newEntry
    AdjacentToAnySet = False
    for consideredSet in consideredSets:
        consideredSetFrozen = deepcopy(consideredSet)
        for (ys, xs) in consideredSetFrozen:
            if abs(y - ys) + abs(x - xs) <= 1:
                if len(consideredSet) < MaxCapacity:
                    consideredSet.add(newEntry)
                AdjacentToAnySet = True
    if not AdjacentToAnySet and CanCreateNewSet:
        consideredSets.append(set([newEntry]))

#Extract new rules from the observations by looking only for observed changes and prediction-observation mismatches
def _Observe(Time, FocusSet, RuleEvidence, oldworld, action, newworld, oldrules, oldnegrules, predictedworld=None):
    newrules = deepcopy(oldrules)
    newnegrules = deepcopy(oldnegrules)
    changesets = []
    valuecount, valuecount_remembered = (dict([]), dict([]))
    #Keep track of cell type counts
    for y in range(height):
        for x in range(width):
            val = oldworld[BOARD][y][x]
            if val not in valuecount:
                valuecount[val] = 1
            else:
                valuecount[val] += 1
    #Update FocusSet based on unique values and unique changing values
    for y in range(height):
        for x in range(width):
            if not _IsPresentlyObserved(Time, newworld, y, x) and not _IsPresentlyObserved(Time-1, newworld, y, x):
                continue
            val = oldworld[BOARD][y][x]
            if valuecount[val] == 1 and val not in FocusSet:
                FocusSet[val] = 0
            if oldworld[BOARD][y][x] != newworld[BOARD][y][x]:
                if valuecount[val] == 1: #unique
                    if val not in FocusSet:
                        FocusSet[val] = 1
                    else:
                        FocusSet[val] += 1
    #Add change sets
    changesets = []
    for y in range(height):
        for x in range(width):
            if not _IsPresentlyObserved(Time, newworld, y, x) and not _IsPresentlyObserved(Time-1, newworld, y, x):
                continue
            if oldworld[BOARD][y][x] != newworld[BOARD][y][x] and oldworld[BOARD][y][x] != '.':
                _AddToAdjacentSet(changesets, (y, x), MaxCapacity=3, CanCreateNewSet=True)
    changesetslen = len(changesets) #the length of change set only
    changeset0len = 0
    if changesetslen > 0:
        changeset0len = len(changesets[0]) #temporary fix: 3 things need to have changed at least to allow for the more complex rules to be induced
    #Add prediction mismatch entries to adjacent change set entry
    for y in range(height):
        for x in range(width):
            if not _IsPresentlyObserved(Time, newworld, y, x) and not _IsPresentlyObserved(Time-1, newworld, y, x):
                continue
            if predictedworld and predictedworld[BOARD][y][x] != newworld[BOARD][y][x] and oldworld[BOARD][y][x] != '.':
                _AddToAdjacentSet(changesets, (y,x), MaxCapacity=2, CanCreateNewSet=True)
    #if there was a change next to a focus set element (spatial dependency)
    chgsets = deepcopy(changesets)
    for changeset in chgsets:
        for (y,x) in changeset:
            if (action == left or action == right) and x>0 and newworld[BOARD][y][x-1] in FocusSet and oldworld[BOARD][y][x-1] != '.':
                _AddToAdjacentSet(changesets, (y,x-1), MaxCapacity=3, CanCreateNewSet=False)
            if (action == left or action == right) and x<width-1 and newworld[BOARD][y][x+1]  in FocusSet and oldworld[BOARD][y][x+1] != '.':
                _AddToAdjacentSet(changesets, (y,x+1), MaxCapacity=3, CanCreateNewSet=False)
            if (action == up or action == down or action == drop) and y>0 and newworld[BOARD][y-1][x]  in FocusSet and oldworld[BOARD][y-1][x] != '.':
                _AddToAdjacentSet(changesets, (y-1,x), MaxCapacity=3, CanCreateNewSet=False)
            if (action == up or action == down) and y<height-1 and newworld[BOARD][y+1][x]  in FocusSet and oldworld[BOARD][y+1][x] != '.':
                _AddToAdjacentSet(changesets, (y+1,x), MaxCapacity=3, CanCreateNewSet=False)
    #filter test
    if "rosfilter" in sys.argv or getIsROSBridge():
        excluded = []
        for changeset in changesets:
            aboutX = False
            for (y,x) in changeset:
                if oldworld[BOARD][y][x] == 'x' or newworld[BOARD][y][x] == 'x':
                    aboutX = True
            if not aboutX or len(changeset) != 2:
                excluded.append(changeset)
        for changeset in excluded:
            if changeset in changesets:
                changesets.remove(changeset)
    print(changesets)
    #Build rules based on changes and prediction-observation mismatches
    for changeset in changesets:
        for (y1_abs,x1_abs) in changeset:
            action_values_precondition = [action, oldworld[VALUES][1:]]
            preconditions = []
            midstate = None
            CONTINUE = False
            for (y2_abs, x2_abs) in changeset:
                (y2_rel, x2_rel) = (y2_abs-y1_abs, x2_abs-x1_abs)
                condition = (y2_rel, x2_rel, oldworld[BOARD][y2_abs][x2_abs])
                if oldworld[BOARD][y2_abs][x2_abs] == '.': #NECESSARY FOR EPISODE RESET ONLY
                    CONTINUE = True
                    break
                if Hypothesis_ValidCondition(condition):
                    preconditions.append(condition)
                    if y2_rel == 0 and x2_rel == 0:
                        if oldworld[BOARD][y2_abs][x2_abs] == newworld[BOARD][y1_abs][x1_abs]:
                            CONTINUE = True
                            break
            if CONTINUE:
                continue
            preconditions = sorted(preconditions)
            for pr in preconditions:
                action_values_precondition.append(pr)
            rule = (tuple(action_values_precondition), (0, 0, newworld[BOARD][y1_abs][x1_abs], tuple([newworld[VALUES][0]-oldworld[VALUES][0]] + list(newworld[VALUES][1:]))))
            if len(preconditions) >= 2 and (changeset0len == 3 or len(preconditions) <= 2):
                RuleEvidence, newrules = Hypothesis_Confirmed(FocusSet, RuleEvidence, newrules, newnegrules, rule)
        break #speedup
    #if rule conditions are only partly met or the predicted outcome is different than observed, build a specialized rule which has the precondition and conclusion corrected!
    max_focus = None
    if len(FocusSet) > 0:
        max_focus = max(FocusSet, key=lambda k: FocusSet[k])
    (positionscores, highesthighscore) = _MatchHypotheses(FocusSet, oldworld, action, newrules)
    for y in range(height):
        for x in range(width):
            if (y,x) not in positionscores:
                continue
            if not _IsPresentlyObserved(Time, newworld, y, x) and oldworld[BOARD][y][x] != max_focus and not (newworld[BOARD][y][x] == '.' and oldworld[BOARD][y][x] != '.'):
                continue
            scores, highscore, rule = positionscores[(y,x)]
            #for rule in oldrules:
            if _RuleApplicable(scores, highscore, highesthighscore, rule):
                if rule[1][2] != newworld[BOARD][y][x]:
                    (precondition, consequence) = rule
                    action_score_and_preconditions = list(precondition)
                    values = action_score_and_preconditions[1]
                    corrected_preconditions = []
                    CONTINUE = False
                    has_focus_set_condition = False #TODO!!!
                    for (y_rel, x_rel, requiredstate) in action_score_and_preconditions[2:]:
                        if y+y_rel >= height or y+y_rel < 0 or x+x_rel >= width or x+x_rel < 0:
                            CONTINUE = True
                            break
                        if oldworld[BOARD][y+y_rel][x+x_rel] == max_focus:
                            has_focus_set_condition = True
                        if oldworld[BOARD][y+y_rel][x+x_rel] == ".":
                            CONTINUE = True
                            break
                        corrected_preconditions.append((y_rel, x_rel, oldworld[BOARD][y+y_rel][x+x_rel]))
                    corrected_preconditions = sorted(corrected_preconditions)
                    if CONTINUE or not has_focus_set_condition:
                        continue
                    rule_new = (tuple([action_score_and_preconditions[0], action_score_and_preconditions[1]]
                               + corrected_preconditions), tuple([rule[1][0], rule[1][1], newworld[BOARD][y][x], tuple([newworld[VALUES][0]-oldworld[VALUES][0]] + list(newworld[VALUES][1:]))]))
                    #print("RULE CORRECTION ", y, x, loc, worldchange); Prettyprint_rule(rule); Prettyprint_rule(rule_new)
                    RuleEvidence, newrules = Hypothesis_Confirmed(FocusSet, RuleEvidence, newrules, newnegrules, rule_new)
                    break
    #Crisp match: Add negative evidence for rules which prediction contradicts observation (in a classical AIRIS implementation restricted to deterministic worlds: this part would remove contradicting rules from the rule set and would ensure they can't be re-induced)
    for y in range(height):
        for x in range(width):
            if not _IsPresentlyObserved(Time, newworld, y, x) and oldworld[BOARD][y][x] != max_focus and not (newworld[BOARD][y][x] == '.' and oldworld[BOARD][y][x] != '.'):
                continue
            for rule in oldrules: #find rules which don't work, and add negative evidence for them (classical AIRIS: remove them and add them to newnegrules)
                (precondition, consequence) = rule
                action_score_and_preconditions = list(precondition)
                values = action_score_and_preconditions[1]
                CONTINUE = False
                if action_score_and_preconditions[0] != action: #rule did not apply
                    continue
                for i in range(len(values)):
                    if values[i] != oldworld[VALUES][i+1]: #value didn't match, rule did not apply
                        CONTINUE = True
                        break
                for (y_rel, x_rel, requiredstate) in action_score_and_preconditions[2:]:
                    if y+y_rel >= height or y+y_rel < 0 or x+x_rel >= width or x+x_rel < 0:
                        CONTINUE = True
                        break
                    if oldworld[BOARD][y+y_rel][x+x_rel] != requiredstate: 
                        CONTINUE = True
                        break
                if CONTINUE:
                    continue
                if rule[1][3][0] != newworld[VALUES][0] - oldworld[VALUES][0]:
                    RuleEvidence, newrules, newnegrules = Hypothesis_Contradicted(RuleEvidence, newrules, newnegrules, rule) #score increase did not happen
                    continue
                for k in range(1, len(rule[1][3])): #wrong value prediction
                    if rule[1][3][k] != newworld[VALUES][k]:
                        RuleEvidence, newrules, newnegrules = Hypothesis_Contradicted(RuleEvidence, newrules, newnegrules, rule)
                        CONTINUE = True
                        break
                if CONTINUE:
                    continue
                if rule[1][2] != newworld[BOARD][y][x]:
                    RuleEvidence, newrules, newnegrules = Hypothesis_Contradicted(RuleEvidence, newrules, newnegrules, rule)
    return FocusSet, RuleEvidence, newrules, newnegrules

#Match hypotheses (rules) preconditions to the world, calculating how AIRIS-confident the prediction would be:
def _MatchHypotheses(FocusSet, oldworld, action, rules):
    positionscores = dict([])
    highesthighscore = 0.0
    AttendPositions = set([])
    for y in range(height):
        for x in range(width):
            if oldworld[BOARD][y][x] in FocusSet:
                AttendPositions.add((y,x))
                for rule in rules:
                    (precondition, consequence) = rule
                    action_score_and_preconditions = list(precondition)
                    for (y_rel, x_rel, requiredstate) in action_score_and_preconditions[2:]:
                        AttendPositions.add((y+y_rel, x+x_rel))
    for y in range(height):
        for x in range(width):
            if (y,x) not in AttendPositions:
                continue
            scores = dict([])
            positionscores[(y,x)] = scores
            highscore = 0.0
            highscorerule = None
            for rule in rules:
                (precondition, consequence) = rule
                action_score_and_preconditions = list(precondition)
                values = action_score_and_preconditions[1]
                if action_score_and_preconditions[0] == action:
                    scores[rule] = 0.0
                else:
                    continue
                CONTINUE = False
                for i in range(len(values)):
                    if values[i] != oldworld[VALUES][i+1]:
                        CONTINUE = True
                if CONTINUE:
                    continue
                for (y_rel, x_rel, requiredstate) in action_score_and_preconditions[2:]:
                    if y+y_rel >= height or y+y_rel < 0 or x+x_rel >= width or x+x_rel < 0:
                        CONTINUE = True
                        break
                    if oldworld[BOARD][y+y_rel][x+x_rel] == requiredstate:
                        scores[rule] += 1.0
                if CONTINUE:
                    continue
                scores[rule] /= (len(precondition)-2)
                if scores[rule] > 0.0 and (scores[rule] > highscore or (scores[rule] == highscore and highscorerule is not None and len(rule[0]) > len(highscorerule[0]))):
                    highscore = scores.get(rule, 0.0)
                    highscorerule = rule
            positionscores[(y,x)] = (scores, highscore, highscorerule)
            if highscore > highesthighscore:
                highesthighscore = highscore
    return (positionscores, highesthighscore)

#Whether a rule is applicable: only if it matches better than not at all, and as well as the best matching rule
def _RuleApplicable(scores, highscore, highesthighscore, rule):
    if highscore > 0.0 and scores.get(rule, 0.0) == highesthighscore:
        return True
    return False

#Print score value taking its semantics regarding its value range semantics for planning into account
def NACE_PrintScore(score):
    if score >= 0.0 and score <= 1.0:
        print("certainty:", score)
    else:
        print("desired: True")