Folder configuration parameter added

avatar/evaluation.py

@@ -2,6 +2,7 @@
 import os
 import argparse
 from util.playout import readVariantFile
+from conf.settings import DATA_PATH
 
 WORK_PATH = os.path.abspath(os.getcwd())
 
@@ -41,15 +42,25 @@ def intersection(lst1, lst2):
     return list(set(ls1) & set(ls2))
 
 def load_data(system, suffix, job, strategy=None):
-    f_train = os.path.join(WORK_PATH, "data", "variants", system + "_train.txt")
-    f_test = os.path.join(WORK_PATH, "data", "variants", system + "_test.txt")
-    f_pop = os.path.join(WORK_PATH, "data", "variants", system + "_pop.txt")
-    f_eval = os.path.join(WORK_PATH, "data", "avatar", "train_data", system + "_eval.txt")
 
-    if strategy is None:
-        f_gan = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + ".txt")
+    if DATA_PATH is None:
+        f_train = os.path.join(WORK_PATH, "data", "variants", system + "_train.txt")
+        f_test = os.path.join(WORK_PATH, "data", "variants", system + "_test.txt")
+        f_pop = os.path.join(WORK_PATH, "data", "variants", system + "_pop.txt")
+        f_eval = os.path.join(WORK_PATH, "data", "avatar", "train_data", system + "_eval.txt")
+        if strategy is None:
+            f_gan = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + ".txt")
+        else:
+            f_gan = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + ".txt")
     else:
-        f_gan = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + ".txt")
+        f_train = os.path.join(DATA_PATH, "variants", system + "_train.txt")
+        f_test = os.path.join(DATA_PATH, "variants", system + "_test.txt")
+        f_pop = os.path.join(DATA_PATH, "variants", system + "_pop.txt")
+        f_eval = os.path.join(DATA_PATH, "avatar", "train_data", system + "_eval.txt")
+        if strategy is None:
+            f_gan = os.path.join(DATA_PATH, "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + ".txt")
+        else:
+            f_gan = os.path.join(DATA_PATH, "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + ".txt")
 
 
     train = readVariantFile(f_train, unique=False)

avatar/generalization.py

@@ -9,6 +9,8 @@
 from pm4py.evaluation.replay_fitness import factory as replay_factory
 from pm4py.evaluation.precision import factory as precision_factory
 
+from conf.settings import DATA_PATH
+
 WORK_PATH = os.path.abspath(os.getcwd())
 
 def readFile(f_name1, f_name2, unique=False):
@@ -95,12 +97,19 @@ def convertToCsv(traces, to_path):
     pn = args.pn
     strategy = args.strategy
 
-    train_file = os.path.join(WORK_PATH, "data", "variants", system + "_train.txt")
-    gen_file = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + ".txt")
-    csv_file = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + "_generalization.csv")
-    xes_file = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + "_generalization.xes")
+    if DATA_PATH is None:
+        train_file = os.path.join(WORK_PATH, "data", "variants", system + "_train.txt")
+        gen_file = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + ".txt")
+        csv_file = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + "_generalization.csv")
+        xes_file = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + "_generalization.xes")
+        pn_file = os.path.join(WORK_PATH, "data", "pns", system, pn)
+    else:
+        train_file = os.path.join(DATA_PATH, "variants", system + "_train.txt")
+        gen_file = os.path.join(DATA_PATH, "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + ".txt")
+        csv_file = os.path.join(DATA_PATH, "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + "_generalization.csv")
+        xes_file = os.path.join(DATA_PATH, "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_" + strategy + "_generalization.xes")
+        pn_file = os.path.join(DATA_PATH, "pns", system, pn)
 
-    pn_file = os.path.join(WORK_PATH, "data", "pns", system, pn)
 
     """ READ FILES AND CONVERT TO XES """
     traces = readFile(train_file,gen_file, unique=True)

avatar/relgan/run.py

@@ -6,6 +6,7 @@
 from avatar.relgan.utils.text_process import text_precess
 from avatar.relgan.utils.utils import pp
 import avatar.relgan.models as models
+from conf.settings import DATA_PATH
 
 WORK_PATH = os.path.abspath(os.getcwd())
 
@@ -62,8 +63,14 @@ def main(given_args=None):
     parser.add_argument('--gen-emb-dim', default=32, type=int, help="generator embedding dimension")
     parser.add_argument('--dis-emb-dim', default=64, type=int, help="TOTAL discriminator embedding dimension")
     parser.add_argument('--num-rep', default=64, type=int, help="number of discriminator embedded representations")
+
+    if DATA_PATH is None:
+        def_dir = os.path.join(WORK_PATH, "data", "avatar", "sgans")
+    else:
+        def_dir = os.path.join(DATA_PATH, "avatar", "sgans")
+
     parser.add_argument('--data-dir',
-                        default=os.path.join(WORK_PATH, "data", "avatar", "sgans"), ## Refer her to where data is stored. Such as /data/julian/data/relgan/data
+                        default=def_dir,
                         type=str,
                         help='Where data data is stored')
 
@@ -78,7 +85,10 @@ def main(given_args=None):
 
     #data_file = os.path.join(args.data_dir, "..", "train_data", '{}.txt'.format(args.dataset))
 
-    seq_vocab_file = os.path.join(WORK_PATH, "data", "variants", str(args.dataset) + "_train.txt")
+    if DATA_PATH is None:
+        seq_vocab_file = os.path.join(WORK_PATH, "data", "variants", str(args.dataset) + "_train.txt")
+    else:
+        seq_vocab_file = os.path.join(DATA_PATH, "variants", str(args.dataset) + "_train.txt")
 
     if args.dataset == 'pb_system_4_1_10':
         args.batch_size = 32

avatar/sampling.py

@@ -6,6 +6,7 @@
 from avatar.util.LoadRelgan import LoadRelgan
 from avatar.util.MHGAN import MHGAN
 from avatar.util.util import writeToFile, readTraces
+from conf.settings import DATA_PATH
 
 WORK_PATH = os.path.abspath(os.getcwd())
 
@@ -54,16 +55,24 @@
         print("****** SAMPLE FOR SUFFIX ", suffix, " ******")
         relgan = LoadRelgan(system=system, suffix=suffix, job=job)
 
-        f_out = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_naive.txt")
+        if DATA_PATH is None:
+            f_out = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_naive.txt")
+        else:
+            f_out = os.path.join(DATA_PATH, "avatar", "variants",system + "_relgan_" + str(suffix) + "_j" + str(job) + "_naive.txt")
+
         print("Start NAIVE SAMPLING")
         gen_samples = relgan.generate(n_samples=n_samples)
         print("Generated samples - shape:", gen_samples.shape)
         print("Writing to file", f_out)
         writeToFile(relgan, f_out, gen_samples)
 
     elif strategy == "mh":
-        eval_path = os.path.join(WORK_PATH, "data", "avatar", "train_data", system + "_eval.txt")
-        f_out = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_mh.txt")
+        if DATA_PATH is None:
+            eval_path = os.path.join(WORK_PATH, "data", "avatar", "train_data", system + "_eval.txt")
+            f_out = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_mh.txt")
+        else:
+            eval_path = os.path.join(DATA_PATH, "avatar", "train_data", system + "_eval.txt")
+            f_out = os.path.join(DATA_PATH, "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job) + "_mh.txt")
 
         tf.reset_default_graph()
         print("****** SAMPLE FOR SUFFIX ", suffix, " ******")

avatar/training.py

@@ -9,19 +9,29 @@
 from avatar.util.LoadRelgan import LoadRelgan
 from avatar.evaluation import load_data, evaluate
 from avatar.util.util import writeToFile
+from conf.settings import DATA_PATH
 
 WORK_PATH = os.path.abspath(os.getcwd())
 
 def find_num_sentences(system):
-    file = os.path.join(WORK_PATH, "data", "avatar", "train_data", system + ".txt")
+    if DATA_PATH is None:
+        file = os.path.join(WORK_PATH, "data", "avatar", "train_data", system + ".txt")
+    else:
+        file = os.path.join(DATA_PATH, "avatar", "train_data", system + ".txt")
+
     return str(sum(1 for _ in open(file)))
 
 def split_train_eval(system, ratio=0.1):
     print("*** Splitting System", system, " to train and evaluation with ratio ", str(ratio), "***")
 
-    load_file = os.path.join(WORK_PATH, "data", "variants", str(system) + "_train.txt")
-    train_out = os.path.join(WORK_PATH, "data", "avatar", "train_data", str(system) + ".txt")
-    eval_out = os.path.join(WORK_PATH, "data", "avatar", "train_data", str(system) + "_eval.txt")
+    if DATA_PATH is None:
+        load_file = os.path.join(WORK_PATH, "data", "variants", str(system) + "_train.txt")
+        train_out = os.path.join(WORK_PATH, "data", "avatar", "train_data", str(system) + ".txt")
+        eval_out = os.path.join(WORK_PATH, "data", "avatar", "train_data", str(system) + "_eval.txt")
+    else:
+        load_file = os.path.join(DATA_PATH, "variants", str(system) + "_train.txt")
+        train_out = os.path.join(DATA_PATH, "avatar", "train_data", str(system) + ".txt")
+        eval_out = os.path.join(DATA_PATH, "avatar", "train_data", str(system) + "_eval.txt")
 
     all_sequences = readVariantFile(load_file, unique=False)
     indices = np.arange(0, len(all_sequences))
@@ -100,7 +110,10 @@ def split_train_eval(system, ratio=0.1):
     cwd = os.path.dirname(os.path.abspath(__file__))
     rootdir = cwd + "/.."
 
-    outdir = os.path.join(WORK_PATH, "data", "avatar", "sgans", system, str(job_id))
+    if DATA_PATH is None:
+        outdir = os.path.join(WORK_PATH, "data", "avatar", "sgans", system, str(job_id))
+    else:
+        outdir = os.path.join(DATA_PATH, "avatar", "sgans", system, str(job_id))
 
     args = [
         # Architecture
@@ -166,7 +179,10 @@ def split_train_eval(system, ratio=0.1):
         print("****** SAMPLE FOR SUFFIX ", suffix, " ******")
         relgan = LoadRelgan(system=system, suffix=suffix, job=job_id)
 
-        f_out = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job_id) + ".txt")
+        if DATA_PATH is None:
+            f_out = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job_id) + ".txt")
+        else:
+            f_out = os.path.join(DATA_PATH, "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job_id) + ".txt")
         print("Start sampling")
         gen_samples = relgan.generate(n_samples=n_samples)
         print(gen_samples.shape)
@@ -178,13 +194,20 @@ def split_train_eval(system, ratio=0.1):
     train, test, eval, pop, _ = load_data(system, "1", job_id)
     results = dict()
     for suffix in ranges:
-        f_gan = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job_id) + ".txt")
+        if DATA_PATH is None:
+            f_gan = os.path.join(WORK_PATH, "data", "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job_id) + ".txt")
+        else:
+            f_gan = os.path.join(DATA_PATH, "avatar", "variants", system + "_relgan_" + str(suffix) + "_j" + str(job_id) + ".txt")
         gan = readVariantFile(f_gan, unique=True)
         res = evaluate(suffix, train, test, eval, pop, gan)
         results[suffix] = res
 
     iw = json.dumps(results)
-    f = open(os.path.join(WORK_PATH, "data", "avatar", "sgans", system, str(job_id), "evaluations_relgan_" + str(system) + "_" + str(job_id) + ".json"), "w")
+    if DATA_PATH is None:
+        f = open(os.path.join(WORK_PATH, "data", "avatar", "sgans", system, str(job_id), "evaluations_relgan_" + str(system) + "_" + str(job_id) + ".json"), "w")
+    else:
+        f = open(os.path.join(DATA_PATH, "avatar", "sgans", system, str(job_id), "evaluations_relgan_" + str(system) + "_" + str(job_id) + ".json"), "w")
+
     f.write(iw)
     f.close()
 
@@ -221,8 +244,12 @@ def split_train_eval(system, ratio=0.1):
         rank_id -= 1
 
     iw = json.dumps(ranks)
-    f = open(os.path.join(WORK_PATH, "data", "avatar", "sgans", system, str(job_id),
-                          "suffix_ranks_relgan_" + str(system) + "_" + str(job_id) + ".json"), "w")
+    if DATA_PATH is None:
+        f = open(os.path.join(WORK_PATH, "data", "avatar", "sgans", system, str(job_id),
+                              "suffix_ranks_relgan_" + str(system) + "_" + str(job_id) + ".json"), "w")
+    else:
+        f = open(os.path.join(DATA_PATH, "avatar", "sgans", system, str(job_id),
+                              "suffix_ranks_relgan_" + str(system) + "_" + str(job_id) + ".json"), "w")
     f.write(iw)
     f.close()
     print("Ranks saved to file.")

avatar/util/LoadRelgan.py

@@ -3,12 +3,16 @@
 import json
 import os
 import time
+from conf.settings import DATA_PATH
 
 WORK_PATH = os.path.abspath(os.getcwd())
 
 class LoadRelgan():
     def __init__(self, system, suffix, job):
-        path = os.path.join(WORK_PATH, "data", "avatar","sgans", system, str(job), "tf_logs", "ckpt")
+        if DATA_PATH is None:
+            path = os.path.join(WORK_PATH, "data", "avatar","sgans", system, str(job), "tf_logs", "ckpt")
+        else:
+            path = os.path.join(DATA_PATH, "avatar", "sgans", system, str(job), "tf_logs", "ckpt")
 
         saver = tf.train.import_meta_graph(os.path.join(path, system + ".adv_model-" + str(suffix) + ".meta"))
 

conf/settings.py

@@ -0,0 +1 @@
+DATA_PATH = "C:\\data"

processes/conformancechecker.py

@@ -6,9 +6,15 @@
 from pm4py.evaluation.replay_fitness import factory as replay_factory
 from pm4py.evaluation.precision import factory as precision_factory
 from pm4py.evaluation.generalization import factory as generalization_factory
+from conf.settings import DATA_PATH
+
+WORK_PATH = os.path.abspath(os.getcwd())
 
 def main(system, miner):
-    log = xes_importer.import_log(os.path.join("data/variants/", str(system) + "_train.xes"))
+    if DATA_PATH is None:
+        log = xes_importer.import_log(os.path.join(WORK_PATH, "data", "variants", str(system) + "_train.xes"))
+    else:
+        log = xes_importer.import_log(os.path.join(DATA_PATH, "variants", str(system) + "_train.xes"))
 
     bestmodel = None
     bestfit = None
@@ -21,9 +27,18 @@ def main(system, miner):
     gen_bestPrec = None
     gen_bestGen = 0
 
-    for file in os.listdir("data/pns/" + str(system) + "/"):
+    if DATA_PATH is None:
+        dir = os.listdir(os.path.join(WORK_PATH, "data", "pns", str(system)))
+    else:
+        dir = os.listdir(os.path.join(DATA_PATH, "pns", str(system)))
+
+    for file in dir:
         if system in file and miner in file:
-            path = os.path.join("data/pns/", str(system), file)
+            if DATA_PATH is None:
+                path = os.path.join(WORK_PATH, "data", "pns", str(system), file)
+            else:
+                path = os.path.join(DATA_PATH, "pns", str(system), file)
+
             print("Checking conformance of file:", path)
 
             net, initial_marking, final_marking = pnml_importer.import_net(path)

processes/playout.py

@@ -5,6 +5,7 @@
 from pm4py.objects.petri.importer import pnml as pnml_importer
 from pm4py.algo.simulation.playout.versions import basic_playout as playout
 from util.playout import readVariantFile, getMaxVariantLength
+from conf.settings import DATA_PATH
 
 WORK_PATH = os.path.abspath(os.getcwd())
 
@@ -60,11 +61,18 @@ def str2bool(v):
     n_traces = int(args.traces)
     eval_only = str2bool(args.eval_only)
 
-    f_pop = os.path.join(WORK_PATH, "data", "variants", str(system) + "_pop.txt")
-    f_train = os.path.join(WORK_PATH, "data", "variants", str(system) + "_train.txt")
-    f_test = os.path.join(WORK_PATH, "data", "variants", str(system) + "_test.txt")
-    f_out = os.path.join(WORK_PATH, "data", "variants", pn + ".txt")
-    f_pn = os.path.join(WORK_PATH, "data", "pns", system, pn)
+    if DATA_PATH is None:
+        f_pop = os.path.join(WORK_PATH, "data", "variants", str(system) + "_pop.txt")
+        f_train = os.path.join(WORK_PATH, "data", "variants", str(system) + "_train.txt")
+        f_test = os.path.join(WORK_PATH, "data", "variants", str(system) + "_test.txt")
+        f_out = os.path.join(WORK_PATH, "data", "variants", pn + ".txt")
+        f_pn = os.path.join(WORK_PATH, "data", "pns", system, pn)
+    else:
+        f_pop = os.path.join(DATA_PATH, "variants", str(system) + "_pop.txt")
+        f_train = os.path.join(DATA_PATH, "variants", str(system) + "_train.txt")
+        f_test = os.path.join(DATA_PATH, "variants", str(system) + "_test.txt")
+        f_out = os.path.join(DATA_PATH, "variants", pn + ".txt")
+        f_pn = os.path.join(DATA_PATH, "pns", system, pn)
 
     seq_len = getMaxVariantLength(f_pop)
     n_decimal = 8

systems/playout/pa_system_10_2.py

@@ -1,4 +1,8 @@
 from util.playout import standard_playout
+import os
+from conf.settings import DATA_PATH
+
+WORK_PATH = os.path.abspath(os.getcwd())
 
 
 if __name__ == "__main__": 

systems/playout/pa_system_11_3.py

@@ -1,4 +1,8 @@
 from util.playout import standard_playout
+import os
+from conf.settings import DATA_PATH
+
+WORK_PATH = os.path.abspath(os.getcwd())
 
 if __name__ == "__main__": 
     pn = "data/systems/pa_system_11_3.pnml"

systems/playout/pa_system_1_5.py

@@ -1,4 +1,8 @@
 from util.playout import standard_playout
+import os
+from conf.settings import DATA_PATH
+
+WORK_PATH = os.path.abspath(os.getcwd())
 
 if __name__ == "__main__": 
     pn = "data/systems/pa_system_1_5.pnml"

systems/playout/pa_system_2_3.py

@@ -1,4 +1,8 @@
 from util.playout import standard_playout
+import os
+from conf.settings import DATA_PATH
+
+WORK_PATH = os.path.abspath(os.getcwd())
 
 if __name__ == "__main__": 
     pn = "data/systems/pa_system_2_3.pnml"

systems/playout/pa_system_4_3.py

@@ -1,4 +1,8 @@
 from util.playout import standard_playout
+import os
+from conf.settings import DATA_PATH
+
+WORK_PATH = os.path.abspath(os.getcwd())
 
 if __name__ == "__main__": 
     pn = "data/systems/pa_system_4_3.pnml"

systems/playout/pa_system_5_5.py

@@ -1,4 +1,8 @@
 from util.playout import standard_playout
+import os
+from conf.settings import DATA_PATH
+
+WORK_PATH = os.path.abspath(os.getcwd())
 
 if __name__ == "__main__": 
     pn = "data/systems/pa_system_5_5.pnml"

systems/playout/pa_system_6_3.py

@@ -1,4 +1,8 @@
 from util.playout import standard_playout
+import os
+from conf.settings import DATA_PATH
+
+WORK_PATH = os.path.abspath(os.getcwd())
 
 if __name__ == "__main__": 
     pn = "data/systems/pa_system_6_3.pnml"