Merge pull request #91 from DES-Lab/alergia-typo

fix sequances

Author: Edi Muškardin

Benchmarking/json_lbt.py

@@ -234,9 +234,9 @@ def corrupt_json(symbols):
 )
 
 
-def generate_dataset(num_sequances):
+def generate_dataset(num_sequences):
     dataset = set()
-    while len(dataset) <= num_sequances:
+    while len(dataset) <= num_sequences:
         ts, tt = generate_random_json(max_depth=2, max_elements=3)
         assert is_valid_json(ts), ts
         dataset.add((tuple(tt), True))
@@ -287,7 +287,7 @@ def validate_string_with_json_parser(json_str, json_parser):
 
 model_learning_dataset = []
 if not use_learned_model:
-    model_learning_dataset = generate_dataset(num_sequances=20000)
+    model_learning_dataset = generate_dataset(num_sequences=20000)
 
     learned_json_model = run_PAPNI(model_learning_dataset, vpa_alphabet)
     # learned_json_model.visualize()
@@ -307,10 +307,10 @@ def validate_string_with_json_parser(json_str, json_parser):
 for _ in range(num_learning_iterations):
     disagreements.clear()
 
-    test_dataset = generate_input_output_data_from_vpa(learned_json_model, num_sequances=10000, max_seq_len=16)
+    test_dataset = generate_input_output_data_from_vpa(learned_json_model, num_sequences=10000, max_seq_len=16)
     print(f"Num well-matched tests: {len([x for x in test_dataset if learned_json_model.is_balanced(x[0])])}")
 
-    num_new_sequances = 0
+    num_new_sequences = 0
 
     for seq, label in test_dataset:
 
@@ -330,9 +330,9 @@ def validate_string_with_json_parser(json_str, json_parser):
             if (seq, label) in model_learning_dataset:
                 model_learning_dataset.remove((seq, label))
             model_learning_dataset.add((seq, not label))
-            num_new_sequances += 1
+            num_new_sequences += 1
 
-    print(f'Added {num_new_sequances} to learning set, total size {len(model_learning_dataset)}')
+    print(f'Added {num_new_sequences} to learning set, total size {len(model_learning_dataset)}')
     learned_json_model = run_PAPNI(model_learning_dataset, vpa_alphabet, print_info=False)
     print(f'Current model size: {learned_json_model.size}')
 

Benchmarking/passive_vpa_vs_rpni.py

@@ -8,7 +8,7 @@
 from statistics import mean, stdev
 
 all_data = dict()
-with open('papni_sequances.pickle', 'rb') as handle:
+with open('papni_sequences.pickle', 'rb') as handle:
     all_data = pickle.load(handle)
 
 
@@ -59,21 +59,21 @@ def evaluate_model(learned_model, test_data):
     return [rpni_model.size, papni_model.size, rpni_error, papni_error]
 
 
-def get_sequances_from_active_sevpa(model):
+def get_sequences_from_active_sevpa(model):
     from aalpy import SUL, run_KV, RandomWordEqOracle, SevpaAlphabet
 
     class CustomSUL(SUL):
         def __init__(self, automatonSUL):
             super(CustomSUL, self).__init__()
             self.sul = automatonSUL
-            self.sequances = []
+            self.sequences = []
 
         def pre(self):
             self.tc = []
             self.sul.pre()
 
         def post(self):
-            self.sequances.append(self.tc)
+            self.sequences.append(self.tc)
             self.sul.post()
 
         def step(self, letter):
@@ -91,7 +91,7 @@ def step(self, letter):
     # eq_oracle = BreadthFirstExplorationEqOracle(vpa_alphabet.get_merged_alphabet(), sul, 7)
     _ = run_KV(alphabet, sul, eq_oracle, automaton_type='vpa', print_level=3)
 
-    return convert_i_o_traces_for_RPNI(sul.sequances)
+    return convert_i_o_traces_for_RPNI(sul.sequences)
 
 
 def split_data_to_learning_and_testing(data, learning_to_test_ratio=0.5):
@@ -104,20 +104,20 @@ def split_data_to_learning_and_testing(data, learning_to_test_ratio=0.5):
     # sorted(data, key=lambda x: len(x[0]))
     shuffle(data)
 
-    learning_sequances, test_sequances = [], []
+    learning_sequences, test_sequences = [], []
 
     l_pos, l_neg = 0, 0
     for seq, label in data:
         if label and l_pos <= num_learning_positive_seq:
-            learning_sequances.append((seq, label))
+            learning_sequences.append((seq, label))
             l_pos += 1
         elif not label and l_neg <= num_learning_negative_seq:
-            learning_sequances.append((seq, label))
+            learning_sequences.append((seq, label))
             l_neg += 1
         else:
-            test_sequances.append((seq, label))
+            test_sequences.append((seq, label))
 
-    return learning_sequances, test_sequances
+    return learning_sequences, test_sequences
 
 
 def run_experiment(experiment_id,
@@ -128,7 +128,7 @@ def run_experiment(experiment_id,
                    learning_to_test_ratio=0.5):
     if random_data_generation:
         data = generate_input_output_data_from_vpa(ground_truth_model,
-                                                   num_sequances=num_of_learning_seq,
+                                                   num_sequences=num_of_learning_seq,
                                                    max_seq_len=max_learning_seq_len,
                                                 )
     else:
@@ -150,7 +150,7 @@ def run_experiment(experiment_id,
         #         wm_negative += 1
         # print(wm_negative)
 
-        # data = get_sequances_from_active_sevpa(ground_truth_model)
+        # data = get_sequences_from_active_sevpa(ground_truth_model)
 
     vpa_alphabet = ground_truth_model.get_input_alphabet()
 

Benchmarking/rpni_papni_memory_footrpint.py

@@ -49,7 +49,7 @@
 # for size in range(5000, 50001, 5000):
 #     print(size)
 #     data = generate_input_output_data_from_vpa(gt,
-#                                                num_sequances=size,
+#                                                num_sequences=size,
 #                                                max_seq_len=randint(6, 30))
 #
 #     y = run_RPNI(data, automaton_type='dfa', print_info=False)

Examples.py

@@ -869,7 +869,7 @@ def rpni_check_model_example():
 
     input_al = model.get_input_alphabet()
 
-    data = generate_input_output_data_from_automata(model, num_sequances=2000,
+    data = generate_input_output_data_from_automata(model, num_sequences=2000,
                                                     min_seq_len=1, max_seq_len=12)
 
     data = convert_i_o_traces_for_RPNI(data)
@@ -1149,7 +1149,7 @@ def passive_vpa_learning_on_lists():
 def passive_vpa_learning_arithmetics():
     from aalpy.learning_algs import run_PAPNI
     from aalpy.utils.BenchmarkVpaModels import gen_arithmetic_data
-    arithmetic_data, vpa_alphabet = gen_arithmetic_data(num_sequances=4000, min_seq_len=2, max_seq_len=10)
+    arithmetic_data, vpa_alphabet = gen_arithmetic_data(num_sequences=4000, min_seq_len=2, max_seq_len=10)
 
     print(f"Alphabet: {vpa_alphabet}")
 
@@ -1168,7 +1168,7 @@ def passive_vpa_learning_on_all_benchmark_models():
 
     for gt in [vpa_L1(), vpa_L12(), vpa_for_odd_parentheses()]:
         vpa_alphabet = gt.input_alphabet
-        data = generate_input_output_data_from_vpa(gt, num_sequances=2000, max_seq_len=16)
+        data = generate_input_output_data_from_vpa(gt, num_sequences=2000, max_seq_len=16)
 
         papni = run_PAPNI(data, vpa_alphabet, algorithm='gsm', print_info=True)
 
@@ -1336,7 +1336,7 @@ def k_tails_example():
                                                    input_alphabet_size=3,
                                                    output_alphabet_size=3)
 
-    data = generate_input_output_data_from_automata(model, num_sequances=2000,
+    data = generate_input_output_data_from_automata(model, num_sequences=2000,
                                                     min_seq_len=1, max_seq_len=12,
                                                     sequance_type='io_traces')
 

aalpy/learning_algs/deterministic_passive/RPNI.py

@@ -80,7 +80,7 @@ def run_PAPNI(data, vpa_alphabet, algorithm='edsm', print_info=True):
 
     assert algorithm in {'gsm', 'classic', 'edsm'}
 
-    # preprocess input sequances to keep track of stack
+    # preprocess input sequences to keep track of stack
     papni_data = []
     for input_seq, label in data:
         # if input sequance is not balanced we do not consider it (it would lead to error state anyway)

aalpy/learning_algs/non_deterministic/OnfsmLstar.py

@@ -28,7 +28,7 @@ def run_non_det_Lstar(alphabet: list, sul: SUL, eq_oracle: Oracle, n_sampling=5,
         n_sampling: number of times that each cell has to be updated. If this number is to low, all-weather condition
             will not hold and learning will not converge to the correct model. (Default value = 50)
 
-        samples: input output sequances provided to learning algorithm. List of ((input sequence), (output sequence)).
+        samples: input output sequences provided to learning algorithm. List of ((input sequence), (output sequence)).
 
         stochastic: if True, non deterministic learning will be performed but probabilities will be added to the
         returned model, making it a stochastic Mealy machine

aalpy/learning_algs/stochastic_passive/FPTA.py

@@ -60,7 +60,7 @@ def create_fpta(data, automaton_type):
 
     for seq in data:
         if automaton_type != 'smm' and seq[0] != root_node.output:
-            print('All sequances passed to Alergia should have the same initial output!')
+            print('All sequences passed to Alergia should have the same initial output!')
             assert False
 
         curr_node = root_node

aalpy/utils/BenchmarkVpaModels.py

@@ -350,7 +350,7 @@ def vpa_for_even_parentheses():
     return vpa
 
 
-def gen_arithmetic_data(num_sequances=3000, min_seq_len=2, max_seq_len=8):
+def gen_arithmetic_data(num_sequences=3000, min_seq_len=2, max_seq_len=8):
     import ast
     from aalpy.base import SUL
     from aalpy.utils import convert_i_o_traces_for_RPNI
@@ -388,7 +388,7 @@ def step(self, letter):
     alphabet = VpaAlphabet(internal_alphabet=['1', '+', ], call_alphabet=['(', ], return_alphabet=[')', ])
     merged_alphabet = alphabet.get_merged_alphabet()
     data = []
-    while len(data) < num_sequances:
+    while len(data) < num_sequences:
         seq = []
         for _ in range(random.randint(min_seq_len, max_seq_len)):
             seq.append(random.choice(merged_alphabet))

aalpy/utils/HelperFunctions.py

@@ -340,14 +340,14 @@ def is_balanced(input_seq, vpa_alphabet):
     return counter == 0
 
 
-def generate_input_output_data_from_automata(model, num_sequances=4000, min_seq_len=1, max_seq_len=16,
+def generate_input_output_data_from_automata(model, num_sequences=4000, min_seq_len=1, max_seq_len=16,
                                              sequance_type='io_traces'):
     assert sequance_type in {'io_traces', 'labeled_sequences'}
 
     alphabet = model.get_input_alphabet()
     dataset = []
 
-    while len(dataset) < num_sequances:
+    while len(dataset) < num_sequences:
         sequance = []
         for _ in range(random.randint(min_seq_len, max_seq_len)):
             sequance.append(random.choice(alphabet))
@@ -363,18 +363,18 @@ def generate_input_output_data_from_automata(model, num_sequances=4000, min_seq_
     return dataset
 
 
-def generate_input_output_data_from_vpa(vpa, num_sequances=1000, max_seq_len=16, max_attempts=None):
+def generate_input_output_data_from_vpa(vpa, num_sequences=1000, max_seq_len=16, max_attempts=None):
     alphabet = vpa.input_alphabet.get_merged_alphabet()
     data_set, in_set = [], set()
 
-    num_nominal_tries = num_sequances // 2
+    num_nominal_tries = num_sequences // 2
     num_generation_attempts = 0
 
-    max_attempts = max_attempts if max_attempts is not None else num_sequances * 50
+    max_attempts = max_attempts if max_attempts is not None else num_sequences * 50
 
-    while len(data_set) < num_sequances:
+    while len(data_set) < num_sequences:
         num_generation_attempts += 1
-        # it can happen that it is not possible to generate num_sequances balanced words of lenght <= max_seq_len
+        # it can happen that it is not possible to generate num_sequences balanced words of lenght <= max_seq_len
         if num_generation_attempts == max_attempts:
             break