Cleanup and updates

- Runner jar
- Optimizations for when calling out to ShadowProver
- Fixed bug in precondition code
- Fixed grapevine domain

pom.xml

@@ -16,17 +16,17 @@
                 <version>2.4</version>
                 <executions>
                     <execution>
-                        <id>sandbox</id>
+                        <id>runner</id>
                         <configuration>
                             <archive>
                                 <manifest>
-                                    <mainClass>org.rairlab.planner.Py4JServer</mainClass>
+                                    <mainClass>org.rairlab.planner.utils.Runner</mainClass>
                                 </manifest>
                             </archive>
                             <descriptorRefs>
                                 <descriptorRef>jar-with-dependencies</descriptorRef>
                             </descriptorRefs>
-                            <finalName>sandbox</finalName>
+                            <finalName>runner</finalName>
                         </configuration>
                         <phase>package</phase>
                         <goals>

src/main/java/org/rairlab/planner/Action.java

@@ -49,7 +49,14 @@ public class Action {
 
         this.freeVariables = freeVariables;
 
-        this.precondition = new And(preconditions.stream().collect(Collectors.toList()));
+        if (preconditions.size() > 1) {
+            this.precondition = new And(preconditions.stream().collect(Collectors.toList()));
+        } else if (preconditions.size() == 1) {
+            this.precondition = preconditions.iterator().next();
+        } else {
+            this.precondition = State.TRUE;
+        }
+
 
         this.weight = preconditions.stream().mapToInt(Formula::getWeight).sum() +
                 additions.stream().mapToInt(Formula::getWeight).sum() +

src/main/java/org/rairlab/planner/Operations.java

@@ -4,11 +4,12 @@ import org.rairlab.planner.utils.Visualizer;
 import org.rairlab.shadow.prover.core.Prover;
 import org.rairlab.shadow.prover.core.ccprovers.CognitiveCalculusProver;
 import org.rairlab.shadow.prover.core.proof.Justification;
-import org.rairlab.shadow.prover.representations.formula.BiConditional;
-import org.rairlab.shadow.prover.representations.formula.Formula;
-
 import org.rairlab.shadow.prover.representations.value.Value;
 import org.rairlab.shadow.prover.representations.value.Variable;
+import org.rairlab.shadow.prover.representations.formula.*;
+import org.rairlab.shadow.prover.representations.value.Constant;
+
+
 import org.rairlab.shadow.prover.utils.CollectionUtils;
 
 import org.rairlab.shadow.prover.utils.Sets;
@@ -16,11 +17,14 @@ import org.apache.commons.lang3.tuple.Pair;
 import org.apache.commons.lang3.tuple.ImmutablePair;
 import org.apache.commons.lang3.tuple.Triple;
 
+import java.util.ArrayList;
+import java.util.HashSet;
 import java.util.List;
 import java.util.Map;
 import java.util.Optional;
 import java.util.Set;
 import java.util.concurrent.*;
+import java.util.stream.Collectors;
 
 /**
  * Created by naveensundarg on 1/13/17.
@@ -29,6 +33,7 @@ public class Operations {
 
     private static boolean DEEP_EQUIVALENCE = false;
     private static boolean THROW_AWAY_EMPTY_BINDINGS = true;
+    private static boolean MONOTONIC = true;
     private static Prover prover;
 
 
@@ -48,40 +53,44 @@ public class Operations {
     }
 
     public static synchronized Optional<Justification> proveCached(Set<Formula> assumptions, Formula goal) {
-
-        // (1) If we've asked to prove this exact goal from assumptions before
-        // then return the previous result
         Pair<Set<Formula>, Formula> inputPair = ImmutablePair.of(assumptions, goal);
-        if (proverCache.containsKey(inputPair)) {
-            return proverCache.get(inputPair);
-        }
 
-        // Iterate through the cache
+        if (MONOTONIC) {
-        for (Map.Entry<Pair<Set<Formula>, Formula>, Optional<Justification>> entry : proverCache.entrySet()) {
+            // (1) If we've asked to prove this exact goal from assumptions before
-            Set<Formula> cachedAssumptions = entry.getKey().getLeft();
+            // then return the previous result
-            Formula cachedGoal = entry.getKey().getRight();
+            if (proverCache.containsKey(inputPair)) {
-            Optional<Justification> optJust = entry.getValue();
+                return proverCache.get(inputPair);
-
-            // (2) Return the cached justification if:
-            // - Goals are the same
-            // - The cached assumptions are a subset of the current ones
-            // - A justification was found
-            if (optJust.isPresent() && cachedGoal.equals(goal) && Sets.subset(cachedAssumptions, assumptions)) {
-                return optJust;
             }
 
-            // (3) Return cached failure if:
+            // Iterate through the cache
-            // - Goals are the same
+            for (Map.Entry<Pair<Set<Formula>, Formula>, Optional<Justification>> entry : proverCache.entrySet()) {
-            // - Assumptions are a subset of cached assumptions
+                Set<Formula> cachedAssumptions = entry.getKey().getLeft();
-            // - No justification was found
+                Formula cachedGoal = entry.getKey().getRight();
-            if (optJust.isEmpty() && cachedGoal.equals(goal) && Sets.subset(assumptions, cachedAssumptions)) {
+                Optional<Justification> optJust = entry.getValue();
-                return optJust;
+
+                // (2) Return the cached justification if:
+                // - Goals are the same
+                // - The cached assumptions are a subset of the current ones
+                // - A justification was found
+                if (optJust.isPresent() && cachedGoal.equals(goal) && Sets.subset(cachedAssumptions, assumptions)) {
+                    return optJust;
+                }
+
+                // (3) Return cached failure if:
+                // - Goals are the same
+                // - Assumptions are a subset of cached assumptions
+                // - No justification was found
+                if (optJust.isEmpty() && cachedGoal.equals(goal) && Sets.subset(assumptions, cachedAssumptions)) {
+                    return optJust;
+                }
             }
-        }
+    }
 
         // Otherwise create a new call to the theorem prover
         Optional<Justification> answer = prover.prove(assumptions, goal);
-        proverCache.put(inputPair, answer);
+        if (MONOTONIC) {
+            proverCache.put(inputPair, answer);
+        }
         return answer;
 
     }
@@ -155,10 +164,9 @@ public class Operations {
         return Optional.of(ans.get().getRight());
     }
 
-
     public static Optional<Set<Pair<State, Action>>> apply(Set<Formula> background, Action action, State state) {
 
-        // Get resulting states from cache if computed before
+        // // Get resulting states from cache if computed before
         if(applyCache.containsKey(Triple.of(background, action, state))){
             Optional<Set<Pair<State, Action>>>  ans  = applyCache.get(Triple.of(background, action, state));
             if(ans.isPresent()){
@@ -169,7 +177,14 @@ public class Operations {
 
         // Ask theorem prover for witnesses that satisfy the precondition
         Set<Formula> givens = Sets.union(background, state.getFormulae());
-        Optional<Set<Map<Variable, Value>>> bindingsOpt = proveAndGetBindingsCached(givens, action.getPrecondition(), action.openVars());
+
+        Formula precondition = action.getPrecondition();
+
+        List<Variable> openVars = action.openVars()
+            .stream()
+            .collect(Collectors.toList());
+
+        Optional<Set<Map<Variable, Value>>> bindingsOpt = proveAndGetBindingsCached(givens, precondition, openVars);
 
         // If not witnesses found, return nothing
         if (!bindingsOpt.isPresent()) {
@@ -191,9 +206,12 @@ public class Operations {
             // newState = (oldState - Deletions(a)) U Additions(a)
             Action groundedAction = action.instantiate(binding);
 
+            Set<Formula> additions = groundedAction.getAdditions();
+            Set<Formula> deletions = groundedAction.getDeletions();
+
             State newState = State.initializeWith(Sets.union(
-                Sets.difference(state.getFormulae(), groundedAction.getDeletions()),
+                Sets.difference(state.getFormulae(), deletions),
-                groundedAction.getAdditions()
+                additions
             ));
 
             // If the state progresses, record it as a possible next state
@@ -203,6 +221,7 @@ public class Operations {
         }
 
         applyCache.put(Triple.of(background, action, state), Optional.of(nextStates));
+
         return Optional.of(nextStates);
 
     }

src/main/java/org/rairlab/planner/State.java

@@ -12,14 +12,17 @@ import java.util.Set;
 public class State {
 
     final Set<Formula> formulae;
+    static Formula TRUE;
     static Formula FALSE;
 
     static{
 
         try {
+            TRUE = Reader.readFormulaFromString("(or P (not P))");
             FALSE = Reader.readFormulaFromString("(and P (not P))");
         } catch (Reader.ParsingException e) {
             e.printStackTrace();
+            TRUE = null;
             FALSE  = null;
         }
 

src/main/java/org/rairlab/planner/search/AStarPlanner.java

@@ -6,6 +6,7 @@ import org.rairlab.planner.State;
 import org.rairlab.planner.Action;
 import org.rairlab.planner.Plan;
 import org.rairlab.planner.Operations;
+import org.rairlab.shadow.prover.representations.value.Value;
 
 import java.util.*;
 import java.util.function.Function;
@@ -67,6 +68,10 @@ public class AStarPlanner {
         comparator.setValue(searchStart, 0);
         search.add(searchStart);
 
+        // For debugging...
+        // Map<State, List<Action>> seq = new HashMap<State, List<Action>>();
+        // seq.put(start, new ArrayList<Action>());
+
         // Current set of plans
         Set<Plan> plansFound = new HashSet<Plan>();
 
@@ -81,7 +86,11 @@ public class AStarPlanner {
             State lastState = currentSearch.getLeft();
             List<Action> previous_actions = currentSearch.getRight();
 
+            // System.out.println("--------------------");
             // System.out.println("Considering state with heuristic: " + comparator.getValue(currentSearch));
+            // System.out.println("Current Plan: " + seq.get(lastState).toString());
+            // System.out.println("Current State: " + lastState.toString());
+            // System.out.println("--------------------");
 
             // Exit loop if we've passed the depth limit
             int currentDepth = previous_actions.size();
@@ -105,6 +114,8 @@ public class AStarPlanner {
 
             // Apply the action to the state and add to the search space
             for (Action action : nonTrivialActions) {
+                // System.out.println("Considering action: " + action.getName());
+
                 Optional<Set<Pair<State, Action>>> optNextStateActionPairs = Operations.apply(background, action, lastState);
 
                 // Ignore actions that aren't applicable
@@ -137,6 +148,10 @@ public class AStarPlanner {
                     int heuristicValue = heuristic.apply(nextState);
                     comparator.setValue(futureSearch, planCost + heuristicValue);
                     search.add(futureSearch);
+
+                    // For debugging...
+                    // seq.put(nextState, next_actions);
+
                 }
             }
         }

src/main/java/org/rairlab/planner/utils/Runner.java

@@ -46,10 +46,11 @@ public final class Runner {
         }
 
         AStarPlanner astarplanner = new AStarPlanner();
-        astarplanner.setK(2);
+        astarplanner.setK(1);
 
         for (PlanningProblem planningProblem : planningProblemList) {
 
+            long start = System.currentTimeMillis();
             Set<Plan> plans = astarplanner.plan(
                 planningProblem.getBackground(),
                 planningProblem.getActions(),
@@ -57,9 +58,11 @@ public final class Runner {
                 planningProblem.getGoal(),
                 ConstantHeuristic::h
             );
+            long end = System.currentTimeMillis();
 
             if(plans.size() > 0) {
                 System.out.println(plans.toString());
+                System.out.println("Time (ms): " + (end - start));
             }
             else {
                 System.out.println("FAILED");

src/main/resources/org/rairlab/planner/problems/epistemic/grapevine.clj

@@ -61,16 +61,17 @@
             ]
             :additions [
                 (Believes! ?a2 (the ?a1))
-                (Believes ?a3 (the ?a1))
+                (Believes! ?a3 (the ?a1))
-                (Believes ?a1 (Believes! ?a2 (the ?a1)))
+                (Believes! ?a1 (Believes! ?a2 (the ?a1)))
-                (Believes ?a1 (Believes! ?a3 (the ?a1)))
+                (Believes! ?a1 (Believes! ?a3 (the ?a1)))
             ]
             :deletions [
                 (not (Believes! ?a2 (the ?a1)))
                 (not (Believes! ?a3 (the ?a1)))
-                (not (Believes ?a1 (Believes! ?a2 (the ?a1))))
+                (not (Believes! ?a1 (Believes! ?a2 (the ?a1))))
-                (not (Believes ?a1 (Believes! ?a3 (the ?a1))))
+                (not (Believes! ?a1 (Believes! ?a3 (the ?a1))))
             ]
+            :cost 2
         })
 
         (define-action share-single [?a1 ?a2 ?a3 ?r] {
@@ -96,6 +97,7 @@
                 (not (Believes! ?a2 (the ?a1)))
                 (not (Believes! ?a1 (Believes! ?a2 (the ?a1))))
             ]
+            :cost 2
         })
 
     ]