correction of optimization and creation of batches

env/default_values.toml

@@ -16,7 +16,7 @@ algorithm_code = "iim 2"
 [mrnn]
 hidden_dim = 10
 learning_rate = 0.01
-iterations = 1000
+iterations = 100
 sequence_length = 7
 
 [explainer]

imputegap/imputation/imputation.py

@@ -53,16 +53,16 @@ def cdrec(ground_truth, contamination, params=None):
 
             :param ground_truth: original time series without contamination
             :param contamination: time series with contamination
-            :param params: [Optional] (truncation_rank, epsilon, iterations) : parameters of the algorithm, if None, default ones are loaded
+            :param params: [Optional] (rank, epsilon, iterations) : parameters of the algorithm, if None, default ones are loaded
 
             :return: imputed_matrix, metrics : all time series with imputation data and their metrics
             """
             if params is not None:
-                truncation_rank, epsilon, iterations = params
+                rank, epsilon, iterations = params
             else:
-                truncation_rank, epsilon, iterations = utils.load_parameters(query="default", algorithm="cdrec")
+                rank, epsilon, iterations = utils.load_parameters(query="default", algorithm="cdrec")
 
-            imputed_matrix = cdrec(contamination=contamination, truncation_rank=truncation_rank, iterations=iterations, epsilon=epsilon)
+            imputed_matrix = cdrec(contamination=contamination, truncation_rank=rank, iterations=iterations, epsilon=epsilon)
 
             metrics = Evaluation(ground_truth, imputed_matrix, contamination).metrics_computation()
 

imputegap/manager/manager.py

@@ -9,15 +9,14 @@
 
 class TimeSeries:
 
-    def __init__(self, data=None, normalization=None):
+    def __init__(self, data=None, normalization=None, limitation_values=None):
         """
         :param ts : Original time series without alteration (ground-truth)
         :param contaminated_ts : time series after contamination
         :param imputation : time series after reconstruction of the missing data
-        :param optimal_params : optimal parameters found for a specific algorithm and time series dataset
-        :param explainer : result of the explainer algorithm to explain the imputation of the time series dataset
+        :param limitation_values : limitation of the maximum number of values by series (computation limitation) | default None
         """
-        self.ts = self.load_timeseries(data, normalization)
+        self.ts = self.load_timeseries(data, normalization, limitation_values)
         self.ts_contaminate = None
         self.ts_imputation = None
         self.metrics = []
@@ -26,14 +25,15 @@ def __init__(self, data=None, normalization=None):
 
 
 
-    def load_timeseries(self, data=None, normalization=None):
+    def load_timeseries(self, data=None, normalization=None, limitation=None):
         """
         Load timeseries manager from file
         FORMAT : (Values,Series), values are seperated by space et series by \n
         @author Quentin Nater
 
         :param filename: path of the time series dataset
         :param normalization : [OPTIONAL] choice of normalization ("z_score" or "min_max")
+        :param limitation : limitation of the maximum number of values by series (computation limitation) | default None
         :return: time series format for imputegap from dataset
         """
 
@@ -43,8 +43,7 @@ def load_timeseries(self, data=None, normalization=None):
 
             if isinstance(data, str):
                 print("\nThe time series has been loaded from " + str(data) + "\n")
-
-                ts = np.genfromtxt(data, delimiter=' ')
+                ts = np.genfromtxt(data, delimiter=' ', max_rows=limitation)
 
             elif isinstance(data, list):
                 print("\nThe time series has been loaded from code ", *data, "\n")

imputegap/manager/utils.py

@@ -44,7 +44,7 @@ def load_parameters(query: str = "default", algorithm: str = "cdrec", dataset: s
     if query == "default":
         filepath = "../env/default_values.toml"
     elif query == "optimal":
-        filepath = "../env/optimal_parameters_"+str(optimizer)+"_"+str(dataset)+"_"+str(algorithm)+".toml"
+        filepath = "../params/optimal_parameters_"+str(optimizer)+"_"+str(dataset)+"_"+str(algorithm)+".toml"
     else:
         print("Query not found for this function ('optimal' or 'default')")
 

imputegap/optimization/bayesian_optimization.py

@@ -23,7 +23,7 @@ def save_optimization(optimal_params, algorithm="cdrec", dataset="", optimizer="
         :param file_name: name of the TOML file to save the results. Default is 'optimization_results.toml'.
         """
         if file_name is None:
-            file_name = "../env/optimal_parameters_" + str(optimizer) + "_" + str(dataset) + "_" + str(algorithm) + ".toml"
+            file_name = "../params/optimal_parameters_" + str(optimizer) + "_" + str(dataset) + "_" + str(algorithm) + ".toml"
 
         if not os.path.exists(file_name):
             file_name = file_name[1:]

imputegap/runner_imputation.py

@@ -4,6 +4,8 @@
 from imputegap.manager.manager import TimeSeries
 import os
 
+from imputegap.optimization.bayesian_optimization import Optimization
+
 
 def display_title(title="Master Thesis", aut="Quentin Nater", lib="ImputeGAP", university="University Fribourg - exascale infolab"):
     print("=" * 100)
@@ -25,6 +27,9 @@ def check_block_size(filename):
     display_title()
 
     filename = "eeg"
+    load = False
+    algo = "cdrec"
+
     file_path = os.path.join("./dataset/", filename + ".txt")
     gap = TimeSeries(data=file_path, normalization="z_score")
 
@@ -34,13 +39,20 @@ def check_block_size(filename):
     gap.plot(title="test", save_path="assets", limitation=0, display=False)
 
     gap.ts_contaminate = Contamination.scenario_mcar(ts=gap.ts, series_impacted=0.4, missing_rate=0.4, block_size=block_size, protection=0.1, use_seed=True, seed=42)
-    gap.print()
+    gap.print(limitation=10)
     gap.plot(ts_type="contamination", title="test", save_path="assets", limitation=plot_limit, display=False)
 
-    gap.optimal_params = utils.load_parameters(query="optimal", algorithm="cdrec")
+    if load:
+        gap.optimal_params = utils.load_parameters(query="optimal", algorithm=algo)
+    else:
+        optimal_params, yi = Optimization.Bayesian.bayesian_optimization(ground_truth=gap.ts,
+                                                                         contamination=gap.ts_contaminate,
+                                                                         algorithm=algo, n_calls=3)
+        gap.optimal_params = tuple(optimal_params.values())
+        print("\nOptical Params : ", gap.optimal_params)
 
     gap.ts_imputation, gap.metrics = Imputation.MR.cdrec(ground_truth=gap.ts, contamination=gap.ts_contaminate, params=gap.optimal_params)
-    gap.print()
+    gap.print(limitation=10)
     gap.print_results()
 
     gap.plot(ts_type="imputation", title="test", save_path="assets", limitation=plot_limit, display=False)

imputegap/runner_optimization.py

@@ -35,12 +35,12 @@ def check_block_size(filename):
         gap.plot(title="test", save_path="assets", limitation=6, display=False)
 
         gap.ts_contaminate = Contamination.scenario_mcar(ts=gap.ts, series_impacted=0.4, missing_rate=0.4, block_size=block_size, protection=0.1, use_seed=True, seed=42)
-        gap.print()
+        gap.print(limitation=5)
         gap.plot(ts_type="contamination", title="test", save_path="assets", limitation=3, display=False)
 
         for algo in ["cdrec", "stmvl", "iim", "mrnn"]:
             print("RUN OPTIMIZATION FOR : ", algo, "... with ", filename, "...")
-            optimal_params, yi = Optimization.Bayesian.bayesian_optimization(ground_truth=gap.ts, contamination=gap.ts_contaminate, algorithm=algo)
+            optimal_params, yi = Optimization.Bayesian.bayesian_optimization(ground_truth=gap.ts, contamination=gap.ts_contaminate, algorithm=algo, n_calls=100)
             print("\nOptical Params : ", optimal_params)
             print("\nyi : ", yi, "\n")
             Optimization.save_optimization(optimal_params=optimal_params, algorithm=algo, dataset=filename, optimizer="b")

tests/test_opti_bayesian_cdrec.py

@@ -22,7 +22,7 @@ def test_optimization_bayesian_cdrec(self):
 
         ts_contaminated = Contamination.scenario_mcar(ts=gap.ts, series_impacted=0.4, missing_rate=0.4, block_size=2, protection=0.1, use_seed=True, seed=42)
 
-        optimal_params, yi = Optimization.Bayesian.bayesian_optimization(ground_truth=gap.ts, contamination=ts_contaminated, algorithm=algorithm)
+        optimal_params, yi = Optimization.Bayesian.bayesian_optimization(ground_truth=gap.ts, contamination=ts_contaminated, algorithm=algorithm, n_calls=3)
 
         print("\nOptimization done successfully... ")
         print("\n", optimal_params, "\n")

tests/test_opti_bayesian_iim.py

@@ -25,7 +25,7 @@ def test_optimization_bayesian_iim(self):
 
         optimal_params, yi = Optimization.Bayesian.bayesian_optimization(ground_truth=gap.ts,
                                                                          contamination=ts_contaminated,
-                                                                         algorithm=algorithm)
+                                                                         algorithm=algorithm, n_calls=3)
 
         print("\nOptimization done successfully... ")
         print("\n", optimal_params, "\n")

tests/test_opti_bayesian_mrnn.py

@@ -24,7 +24,7 @@ def test_optimization_bayesian_mrnn(self):
 
         optimal_params, yi = Optimization.Bayesian.bayesian_optimization(ground_truth=gap.ts,
                                                                          contamination=ts_contaminated,
-                                                                         algorithm=algorithm)
+                                                                         algorithm=algorithm, n_calls=3)
 
         print("\nOptimization done successfully... ")
         print("\n", optimal_params, "\n")

tests/test_opti_bayesian_stmvl.py

@@ -25,7 +25,7 @@ def test_optimization_bayesian_stmvl(self):
 
         optimal_params, yi = Optimization.Bayesian.bayesian_optimization(ground_truth=gap.ts,
                                                                          contamination=ts_contaminated,
-                                                                         algorithm=algorithm)
+                                                                         algorithm=algorithm, n_calls=3)
 
         print("\nOptimization done successfully... ")
         print("\n", optimal_params, "\n")