Merge pull request #79 from DES-Lab/gsm-dev

Resolve Minor GSM issues

Author: Edi Muškardin

DotModels/car_alarm.dot

@@ -5,22 +5,18 @@ q3_locked_closed [label="A"];
 q5_unlocked_closed [label="N"];
 q6_unlocked_open [label="N"];
 q7_locked_open [label="N"];
-q4_faulty [label="N"];
-q1_locked_closed -> q2_locked_open  [label="d:1"];
-q1_locked_closed -> q5_unlocked_closed  [label="l:1"];
-q2_locked_open -> q3_locked_closed  [label="d:1"];
-q2_locked_open -> q6_unlocked_open  [label="l:1"];
-q3_locked_closed -> q2_locked_open  [label="d:1"];
-q3_locked_closed -> q5_unlocked_closed  [label="l:1"];
-q5_unlocked_closed -> q6_unlocked_open  [label="d:1"];
-q5_unlocked_closed -> q1_locked_closed  [label="l:1"];
-q6_unlocked_open -> q5_unlocked_closed  [label="d:1"];
-q6_unlocked_open -> q7_locked_open  [label="l:1"];
-q7_locked_open -> q4_faulty  [label="d:1"];
-q7_locked_open -> q6_unlocked_open  [label="l:1"];
-q4_faulty -> q2_locked_open  [label="d:0.9"];
-q4_faulty -> q7_locked_open  [label="d:0.1"];
-q4_faulty -> q5_unlocked_closed  [label="l:1"];
+q1_locked_closed -> q2_locked_open  [label="d"];
+q1_locked_closed -> q5_unlocked_closed  [label="l"];
+q2_locked_open -> q3_locked_closed  [label="d"];
+q2_locked_open -> q6_unlocked_open  [label="l"];
+q3_locked_closed -> q2_locked_open  [label="d"];
+q3_locked_closed -> q5_unlocked_closed  [label="l"];
+q5_unlocked_closed -> q6_unlocked_open  [label="d"];
+q5_unlocked_closed -> q1_locked_closed  [label="l"];
+q6_unlocked_open -> q5_unlocked_closed  [label="d"];
+q6_unlocked_open -> q7_locked_open  [label="l"];
+q7_locked_open -> q1_locked_closed  [label="d"];
+q7_locked_open -> q6_unlocked_open  [label="l"];
 __start0 [label="", shape=none];
 __start0 -> q1_locked_closed  [label=""];
 }

aalpy/learning_algs/__init__.py

@@ -10,4 +10,5 @@
 from .stochastic_passive.ActiveAleriga import run_active_Alergia
 from .deterministic_passive.RPNI import run_RPNI, run_PAPNI
 from .deterministic_passive.active_RPNI import run_active_RPNI
-from .general_passive.GeneralizedStateMerging import run_GSM
+from .general_passive.GeneralizedStateMerging import run_GSM
+from .general_passive.GsmAlgorithms import run_EDSM

aalpy/learning_algs/general_passive/GeneralizedStateMerging.py

@@ -1,10 +1,9 @@
 import functools
-from bisect import insort
-from typing import Dict, Tuple, Callable, List, Optional
 from collections import deque
+from typing import Dict, Tuple, Callable, List, Optional
 
 from aalpy.learning_algs.general_passive.Node import Node, OutputBehavior, TransitionBehavior, TransitionInfo, \
-    OutputBehaviorRange, TransitionBehaviorRange, intersection_iterator
+    OutputBehaviorRange, TransitionBehaviorRange, intersection_iterator, NodeOrders, unknown_output, detect_data_format
 from aalpy.learning_algs.general_passive.ScoreFunctionsGSM import ScoreCalculation, hoeffding_compatibility
 
 
@@ -54,10 +53,10 @@ def __init__(self, *,
                  depth_first=False):
 
         if output_behavior not in OutputBehaviorRange:
-            raise ValueError(f"invalid output behavior {output_behavior}")
+            raise ValueError(f"invalid output behavior {output_behavior}. should be in {OutputBehaviorRange}")
         self.output_behavior: OutputBehavior = output_behavior
         if transition_behavior not in TransitionBehaviorRange:
-            raise ValueError(f"invalid transition behavior {transition_behavior}")
+            raise ValueError(f"invalid transition behavior {transition_behavior}. should be in {TransitionBehaviorRange}")
         self.transition_behavior: TransitionBehavior = transition_behavior
 
         if score_calc is None:
@@ -70,8 +69,11 @@ def __init__(self, *,
         self.score_calc: ScoreCalculation = score_calc
 
         if node_order is None:
-            node_order = Node.__lt__
-        self.node_order = functools.cmp_to_key(lambda a, b: -1 if node_order(a, b) else 1)
+            node_order = NodeOrders.Default
+        if node_order is NodeOrders.NoCompare or node_order is NodeOrders.Default:
+            self.node_order = node_order
+        else:
+            self.node_order = functools.cmp_to_key(lambda a, b: -1 if node_order(a, b) else 1)
 
         self.pta_preprocessing = pta_preprocessing or (lambda x: x)
         self.postprocessing = postprocessing or (lambda x: x)
@@ -91,15 +93,16 @@ def compute_local_compatibility(self, a: Node, b: Node):
 
     # TODO: make more generic by adding the option to use a different algorithm than red blue
     #  for selecting potential merge candidates. Maybe using inheritance with abstract `run`.
-    def run(self, data, convert=True, instrumentation: Instrumentation = None):
+    def run(self, data, convert=True, instrumentation: Instrumentation=None, data_format=None):
         if instrumentation is None:
             instrumentation = Instrumentation()
         instrumentation.reset(self)
 
-        if isinstance(data, Node):
-            root = data
-        else:
-            root = Node.createPTA(data, self.output_behavior)
+        if data_format is None:
+            data_format = detect_data_format(data)
+        if data_format == "examples" and self.transition_behavior != "deterministic":
+            raise ValueError("learning from examples is not possible for nondeterministic systems")
+        root = Node.createPTA(data, self.output_behavior, data_format)
 
         root = self.pta_preprocessing(root)
         instrumentation.pta_construction_done(root)
@@ -128,7 +131,11 @@ def run(self, data, convert=True, instrumentation: Instrumentation = None):
             # no blue states left -> done
             if len(blue_states) == 0:
                 break
-            blue_states.sort(key=self.node_order)
+            if self.node_order is not NodeOrders.NoCompare:
+                blue_states.sort(key=self.node_order)
+                # red states are always sorted using default order on original prefix
+                if self.node_order is not NodeOrders.Default:
+                    red_states.sort(key=self.node_order)
 
             # loop over blue states
             promotion = False
@@ -139,7 +146,6 @@ def run(self, data, convert=True, instrumentation: Instrumentation = None):
                 # calculate partitions resulting from merges with red states if necessary
                 current_candidates: Dict[Node, Partitioning] = dict()
                 perfect_partitioning = None
-
                 red_state = None
                 for red_state in red_states:
                     partition = partition_candidates.get((red_state, blue_state))
@@ -149,16 +155,16 @@ def run(self, data, convert=True, instrumentation: Instrumentation = None):
                         perfect_partitioning = partition
                         break
                     current_candidates[red_state] = partition
-
                 assert red_state is not None
+
                 # partition with perfect score found: don't consider anything else
                 if perfect_partitioning:
                     partition_candidates = {(red_state, blue_state): perfect_partitioning}
                     break
 
                 # no merge candidates for this blue state -> promote
                 if all(part.score is False for part in current_candidates.values()):
-                    insort(red_states, blue_state, key=self.node_order)
+                    red_states.append(blue_state)
                     instrumentation.log_promote(blue_state)
                     promotion = True
                     break
@@ -176,10 +182,11 @@ def run(self, data, convert=True, instrumentation: Instrumentation = None):
             best_candidate = max(partition_candidates.values(), key=lambda part: part.score)
             for real_node, partition_node in best_candidate.red_mapping.items():
                 real_node.transitions = partition_node.transitions
+                real_node.prefix_access_pair = partition_node.prefix_access_pair
                 for access_pair, t_info in real_node.transition_iterator():
                     if t_info.target not in red_states:
                         t_info.target.predecessor = real_node
-                        t_info.target.prefix_access_pair = access_pair  # not sure whether this is actually required
+                        # t_info.target.prefix_access_pair = access_pair  # not sure whether this is actually required
             instrumentation.log_merge(best_candidate)
             # FUTURE: optimizations for compatibility tests where merges can be orthogonal
             # FUTURE: caching for aggregating compatibility tests
@@ -247,9 +254,13 @@ def update_partition(red_node: Node, blue_node: Optional[Node]) -> Node:
         blue_in_sym, blue_out_sym = blue.prefix_access_pair
         blue_parent.transitions[blue_in_sym][blue_out_sym].target = red
 
+        partition = update_partition(red, None)
+        if self.output_behavior == "moore":
+            partition.resolve_unknown_prefix_output(blue_out_sym)
+
+        # loop over implied merges
         q: deque[Tuple[Node, Node]] = deque([(red, blue)])
         pop = q.pop if self.depth_first else q.popleft
-
         while len(q) != 0:
             red, blue = pop()
             partition = update_partition(red, blue)
@@ -258,10 +269,25 @@ def update_partition(red_node: Node, blue_node: Optional[Node]) -> Node:
                 if self.compute_local_compatibility(partition, blue) is False:
                     return partitioning
 
+            # create implied merges for all common successors
             for in_sym, blue_transitions in blue.transitions.items():
                 partition_transitions = partition.get_or_create_transitions(in_sym)
                 for out_sym, blue_transition in blue_transitions.items():
                     partition_transition = partition_transitions.get(out_sym)
+                    # handle unknown output
+                    if partition_transition is None:
+                        if out_sym is unknown_output and len(partition_transitions) != 0:
+                            assert len(partition_transitions) == 1
+                            partition_transition = list(partition_transitions.values())[0]
+                        if unknown_output in partition_transitions:
+                            assert len(partition_transitions) == 1
+                            partition_transition = partition_transitions.pop(unknown_output)
+                            partition_transitions[out_sym] = partition_transition
+                            # re-hook access pair
+                            succ_part = update_partition(partition_transition.target, None)
+                            if self.output_behavior == "moore" or succ_part.predecessor is red:
+                                succ_part.resolve_unknown_prefix_output(out_sym)
+                    # add pairs
                     if partition_transition is not None:
                         q.append((partition_transition.target, blue_transition.target))
                         partition_transition.count += blue_transition.count
@@ -287,6 +313,7 @@ def run_GSM(data, *,
             depth_first=False,
             instrumentation=None,
             convert=True,
+            data_format=None,
             ):
     """
     TODO
@@ -318,12 +345,14 @@ def run_GSM(data, *,
 
         convert:
 
+        data_format:
+
 
     Returns:
 
 
     """
-    # instantiate the gsm
+    # instantiate gsm
     gsm = GeneralizedStateMerging(
         output_behavior=output_behavior,
         transition_behavior=transition_behavior,
@@ -338,4 +367,4 @@ def run_GSM(data, *,
     )
 
     # run the algorithm
-    return gsm.run(data=data, instrumentation=instrumentation, convert=convert)
+    return gsm.run(data=data, instrumentation=instrumentation, convert=convert, data_format=data_format)

aalpy/learning_algs/general_passive/GsmAlgorithms.py

@@ -0,0 +1,57 @@
+from typing import Dict, Union
+
+from aalpy import DeterministicAutomaton
+from aalpy.learning_algs.general_passive.GeneralizedStateMerging import run_GSM
+from aalpy.learning_algs.general_passive.Instrumentation import ProgressReport
+from aalpy.learning_algs.general_passive.Node import Node
+from aalpy.learning_algs.general_passive.ScoreFunctionsGSM import ScoreCalculation
+from aalpy.utils.HelperFunctions import dfa_from_moore
+
+
+def run_EDSM(data, automaton_type, input_completeness=None, print_info=True) -> Union[DeterministicAutomaton, None]:
+    """
+    Run Evidence Driven State Merging.
+
+    Args:
+        data: sequence of input sequences and corresponding label. Eg. [[(i1,i2,i3, ...), label], ...]
+        automaton_type: either 'dfa', 'mealy', 'moore'. Note that for 'mealy' machine learning, data has to be prefix-closed.
+        input_completeness: either None, 'sink_state', or 'self_loop'. If None, learned model could be input incomplete,
+        sink_state will lead all undefined inputs form some state to the sink state, whereas self_loop will simply create
+        a self loop. In case of Mealy learning output of the added transition will be 'epsilon'.
+        print_info: print learning progress and runtime information
+
+    Returns:
+
+        Model conforming to the data, or None if data is non-deterministic.
+
+    """
+    assert automaton_type in {'dfa', 'mealy', 'moore'}
+    assert input_completeness in {None, 'self_loop', 'sink_state'}
+
+    def EDSM_score(part: Dict[Node, Node]):
+        nr_partitions = len(set(part.values()))
+        nr_merged = len(part)
+        return nr_merged - nr_partitions
+
+    score = ScoreCalculation(score_function=EDSM_score)
+
+    internal_automaton_type = 'moore' if automaton_type != 'mealy' else automaton_type
+
+    learned_model = run_GSM(data, output_behavior=internal_automaton_type,
+                            transition_behavior="deterministic",
+                            score_calc=score, instrumentation=ProgressReport(2))
+
+    if automaton_type == 'dfa':
+        learned_model = dfa_from_moore(learned_model)
+
+    if not learned_model.is_input_complete():
+        if not input_completeness:
+            if print_info:
+                print('Warning: Learned Model is not input complete (inputs not defined for all states). '
+                      'Consider calling .make_input_complete()')
+        else:
+            if print_info:
+                print(f'Learned model was not input complete. Adapting it with {input_completeness} transitions.')
+            learned_model.make_input_complete(input_completeness)
+
+    return learned_model

aalpy/learning_algs/general_passive/Instrumentation.py

@@ -1,5 +1,4 @@
 import time
-from functools import wraps
 from typing import Dict, Optional
 
 from aalpy.learning_algs.general_passive.GeneralizedStateMerging import Instrumentation, Partitioning, \
@@ -8,19 +7,6 @@
 
 
 class ProgressReport(Instrumentation):
-    @staticmethod
-    def min_lvl(lvl):
-        def decorator(fn):
-            @wraps(fn)
-            def wrapper(this, *args, **kw):
-                if this.lvl < lvl:
-                    return
-                fn(this, *args, **kw)
-
-            return wrapper
-
-        return decorator
-
     def __init__(self, lvl):
         super().__init__()
         self.lvl = lvl
@@ -45,10 +31,9 @@ def reset(self, gsm: GeneralizedStateMerging):
 
         self.previous_time = time.time()
 
-    @min_lvl(1)
     def pta_construction_done(self, root):
         print(f'PTA Construction Time: {round(time.time() - self.previous_time, 2)}')
-        if self.lvl != 1:
+        if 1 < self.lvl:
             states = root.get_all_nodes()
             leafs = [state for state in states if len(state.transitions.keys()) == 0]
             depth = [state.get_prefix_length() for state in leafs]
@@ -60,24 +45,21 @@ def pta_construction_done(self, root):
     def print_status(self):
         reset_char = "\33[2K\r"
         print_str = reset_char + f'Current automaton size: {self.nr_red_states}'
-        if self.lvl != 1 and not self.gsm.compatibility_on_futures:
+        if 1 < self.lvl and not self.gsm.compatibility_on_futures:
             print_str += f' Merged: {self.nr_merged_states_total} Remaining: {self.pta_size - self.nr_red_states - self.nr_merged_states_total}'
         print(print_str, end="")
 
-    @min_lvl(1)
     def log_promote(self, node: Node):
         self.log.append(["promote", (node.get_prefix(),)])
         self.nr_red_states += 1
         self.print_status()
 
-    @min_lvl(1)
     def log_merge(self, part: Partitioning):
         self.log.append(["merge", (part.red.get_prefix(), part.blue.get_prefix())])
         self.nr_merged_states_total += len(part.full_mapping) - len(part.red_mapping)
         self.nr_merged_states += 1
         self.print_status()
 
-    @min_lvl(1)
     def learning_done(self, root, red_states):
         print(f'\nLearning Time: {round(time.time() - self.previous_time, 2)}')
         print(f'Learned {len(red_states)} state automaton via {self.nr_merged_states} merges.')