add deepdta

AttentionDTA_BIBM @ d8c8a667

-Subproject commit d8c8a6673c75c5e457ccf5bc0c187b131a06b7a1

DeepDTA @ 2c9cbafd

-Subproject commit 2c9cbafdfb383f2f03bcea4b231b90a072e65b15

baselines/DeepDTA/README.md

+# About DeepDTA: deep drug-target binding affinity prediction
+
+The approach used in this work is the modeling of protein sequences and compound 1D representations (SMILES) with convolutional neural networks (CNNs) to predict the binding affinity value of drug-target pairs.
+
+![Figure](https://github.com/hkmztrk/DeepDTA/blob/master/docs/figures/deepdta.PNG)
+# Installation
+
+## Data
+
+Please see the [readme](https://github.com/hkmztrk/DeepDTA/blob/master/data/README.md) for detailed explanation.
+
+## Requirements
+
+You'll need to install following in order to run the codes.
+
+*  [Python 3.4 <=](https://www.python.org/downloads/)
+*  [Keras 2.x](https://pypi.org/project/Keras/)
+*  [Tensorflow 1.x](https://www.tensorflow.org/install/)
+*  numpy
+*  matplotlib
+
+You have to place "data" folder under "source" directory. 
+
+# Usage
+```
+python run_experiments.py --num_windows 32 \
+                          --seq_window_lengths 8 12 \
+                          --smi_window_lengths 4 8 \
+                          --batch_size 256 \
+                          --num_epoch 100 \
+                          --max_seq_len 1000 \
+                          --max_smi_len 100 \
+                          --dataset_path 'data/kiba/' \
+                          --problem_type 1 \
+                          --log_dir 'logs/'
+
+
+```
+
+
+
+
+
+**For citation:**
+
+```
+@article{ozturk2018deepdta,
+  title={DeepDTA: deep drug--target binding affinity prediction},
+  author={{\"O}zt{\"u}rk, Hakime and {\"O}zg{\"u}r, Arzucan and Ozkirimli, Elif},
+  journal={Bioinformatics},
+  volume={34},
+  number={17},
+  pages={i821--i829},
+  year={2018},
+  publisher={Oxford University Press}
+}
+```

baselines/DeepDTA/source/arguments.py

+import argparse
+import os
+
+
+
+def argparser():
+  parser = argparse.ArgumentParser()
+  # for model
+  parser.add_argument(
+      '--seq_window_lengths',
+      type=int,
+      nargs='+',
+      help='Space seperated list of motif filter lengths. (ex, --window_lengths 4 8 12)'
+  )
+  parser.add_argument(
+      '--smi_window_lengths',
+      type=int,
+      nargs='+',
+      help='Space seperated list of motif filter lengths. (ex, --window_lengths 4 8 12)'
+  )
+  parser.add_argument(
+      '--num_windows',
+      type=int,
+      nargs='+',
+      help='Space seperated list of the number of motif filters corresponding to length list. (ex, --num_windows 100 200 100)'
+  )
+  parser.add_argument(
+      '--num_hidden',
+      type=int,
+      default=0,
+      help='Number of neurons in hidden layer.'
+  )
+  parser.add_argument(
+      '--num_classes',
+      type=int,
+      default=0,
+      help='Number of classes (families).'
+  )
+  parser.add_argument(
+      '--max_seq_len',
+      type=int,
+      default=0,
+      help='Length of input sequences.'
+  )
+  parser.add_argument(
+      '--max_smi_len',
+      type=int,
+      default=0,
+      help='Length of input sequences.'
+  )
+  # for learning
+  parser.add_argument(
+      '--learning_rate',
+      type=float,
+      default=0.001,
+      help='Initial learning rate.'
+  )
+  parser.add_argument(
+      '--num_epoch',
+      type=int,
+      default=100,
+      help='Number of epochs to train.'
+  )
+  parser.add_argument(
+      '--batch_size',
+      type=int,
+      default=256,
+      help='Batch size. Must divide evenly into the dataset sizes.'
+  )
+  parser.add_argument(
+      '--dataset_path',
+      type=str,
+      default='/data/kiba/',
+      help='Directory for input data.'
+  )
+  parser.add_argument(
+      '--problem_type',
+      type=int,
+      default=1,
+      help='Type of the prediction problem (1-4)'
+  )
+  parser.add_argument(
+      '--binary_th',
+      type=float,
+      default=0.0,
+      help='Threshold to split data into binary classes'
+  )
+  parser.add_argument(
+      '--is_log',
+      type=int,
+      default=0,
+      help='use log transformation for Y'
+  )
+  parser.add_argument(
+      '--checkpoint_path',
+      type=str,
+      default='',
+      help='Path to write checkpoint file.'
+  )
+  parser.add_argument(
+      '--log_dir',
+      type=str,
+      default='/tmp',
+      help='Directory for log data.'
+  )
+  parser.add_argument(
+      '--out',
+      type=str,
+      default='/pred',
+      help='Directory for log data.'
+  )
+  parser.add_argument(
+      '--model',
+      type=str,
+      default='/model',
+      help='Directory for log data.'
+  )
+
+
+
+  FLAGS, unparsed = parser.parse_known_args()
+
+  # check validity
+  #assert( len(FLAGS.window_lengths) == len(FLAGS.num_windows) )
+
+  return FLAGS
+
+
+
+
+def logging(msg, FLAGS):
+  fpath = os.path.join( FLAGS.log_dir, "log.txt" )
+  with open( fpath, "a" ) as fw:
+    fw.write("%s\n" % msg)
+  #print(msg)

baselines/DeepDTA/source/datahelper.py

+import sys, re, math, time
+import numpy as np
+import matplotlib.pyplot as plt
+import json
+import pickle
+import collections
+from collections import OrderedDict
+from matplotlib.pyplot import cm
+from spacy import load
+#from keras.preprocessing.sequence import pad_sequences
+
+
+## ######################## ##
+#
+#  Define CHARSET, CHARLEN
+#
+## ######################## ## 
+
+# CHARPROTSET = { 'A': 0, 'C': 1, 'D': 2, 'E': 3, 'F': 4, 'G': 5, 'H': 6, \
+#             'I': 7, 'K': 8, 'L': 9, 'M': 10, 'N': 11, 'P': 12, 'Q': 13, \
+#             'R': 14, 'S': 15, 'T': 16, 'V': 17, 'W': 18, 'Y': 19, 'X': 20, \
+#             'O': 20, 'U': 20,
+#             'B': (2, 11),
+#             'Z': (3, 13),
+#             'J': (7, 9) }
+# CHARPROTLEN = 21
+
+CHARPROTSET = { "A": 1, "C": 2, "B": 3, "E": 4, "D": 5, "G": 6, 
+				"F": 7, "I": 8, "H": 9, "K": 10, "M": 11, "L": 12, 
+				"O": 13, "N": 14, "Q": 15, "P": 16, "S": 17, "R": 18, 
+				"U": 19, "T": 20, "W": 21, 
+				"V": 22, "Y": 23, "X": 24, 
+				"Z": 25 }
+
+CHARPROTLEN = 25
+
+CHARCANSMISET = { "#": 1, "%": 2, ")": 3, "(": 4, "+": 5, "-": 6, 
+			 ".": 7, "1": 8, "0": 9, "3": 10, "2": 11, "5": 12, 
+			 "4": 13, "7": 14, "6": 15, "9": 16, "8": 17, "=": 18, 
+			 "A": 19, "C": 20, "B": 21, "E": 22, "D": 23, "G": 24,
+			 "F": 25, "I": 26, "H": 27, "K": 28, "M": 29, "L": 30, 
+			 "O": 31, "N": 32, "P": 33, "S": 34, "R": 35, "U": 36, 
+			 "T": 37, "W": 38, "V": 39, "Y": 40, "[": 41, "Z": 42, 
+			 "]": 43, "_": 44, "a": 45, "c": 46, "b": 47, "e": 48, 
+			 "d": 49, "g": 50, "f": 51, "i": 52, "h": 53, "m": 54, 
+			 "l": 55, "o": 56, "n": 57, "s": 58, "r": 59, "u": 60,
+			 "t": 61, "y": 62}
+
+CHARCANSMILEN = 62
+
+CHARISOSMISET = {"#": 29, "%": 30, ")": 31, "(": 1, "+": 32, "-": 33, "/": 34, ".": 2, 
+				"1": 35, "0": 3, "3": 36, "2": 4, "5": 37, "4": 5, "7": 38, "6": 6, 
+				"9": 39, "8": 7, "=": 40, "A": 41, "@": 8, "C": 42, "B": 9, "E": 43, 
+				"D": 10, "G": 44, "F": 11, "I": 45, "H": 12, "K": 46, "M": 47, "L": 13, 
+				"O": 48, "N": 14, "P": 15, "S": 49, "R": 16, "U": 50, "T": 17, "W": 51, 
+				"V": 18, "Y": 52, "[": 53, "Z": 19, "]": 54, "\\": 20, "a": 55, "c": 56, 
+				"b": 21, "e": 57, "d": 22, "g": 58, "f": 23, "i": 59, "h": 24, "m": 60, 
+				"l": 25, "o": 61, "n": 26, "s": 62, "r": 27, "u": 63, "t": 28, "y": 64}
+
+CHARISOSMILEN = 64
+
+
+## ######################## ##
+#
+#  Encoding Helpers
+#
+## ######################## ## 
+
+#  Y = -(np.log10(Y/(math.pow(math.e,9))))
+
+def one_hot_smiles(line, MAX_SMI_LEN, smi_ch_ind):
+	X = np.zeros((MAX_SMI_LEN, len(smi_ch_ind))) #+1
+
+	for i, ch in enumerate(line[:MAX_SMI_LEN]):
+		X[i, (smi_ch_ind[ch]-1)] = 1 
+
+	return X #.tolist()
+
+def one_hot_sequence(line, MAX_SEQ_LEN, smi_ch_ind):
+	X = np.zeros((MAX_SEQ_LEN, len(smi_ch_ind))) 
+	for i, ch in enumerate(line[:MAX_SEQ_LEN]):
+		X[i, (smi_ch_ind[ch])-1] = 1
+
+	return X #.tolist()
+
+
+def label_smiles(line, MAX_SMI_LEN, smi_ch_ind):
+	X = np.zeros(MAX_SMI_LEN)
+	for i, ch in enumerate(line[:MAX_SMI_LEN]): #	x, smi_ch_ind, y
+		X[i] = smi_ch_ind[ch]
+
+	return X #.tolist()
+
+def label_sequence(line, MAX_SEQ_LEN, smi_ch_ind):
+	X = np.zeros(MAX_SEQ_LEN)
+
+	for i, ch in enumerate(line[:MAX_SEQ_LEN]):
+		X[i] = smi_ch_ind[ch]
+
+	return X #.tolist()
+
+
+
+## ######################## ##
+#
+#  DATASET Class
+#
+## ######################## ## 
+# works for large dataset
+class DataSet(object):
+  def __init__(self, fpath, setting_no, seqlen, smilen, need_shuffle = False):
+    self.SEQLEN = seqlen
+    self.SMILEN = smilen
+    #self.NCLASSES = n_classes
+    self.charseqset = CHARPROTSET
+    self.charseqset_size = CHARPROTLEN
+
+    self.charsmiset = CHARISOSMISET ###HERE CAN BE EDITED
+    self.charsmiset_size = CHARISOSMILEN
+    self.PROBLEMSET = setting_no 
+
+    # read raw file
+    # self._raw = self.read_sets( FLAGS)
+
+    # iteration flags
+    # self._num_data = len(self._raw)
+
+
+  def read_sets(self, FLAGS): ### fpath should be the dataset folder /kiba/ or /davis/
+    fpath = FLAGS.dataset_path
+    setting_no = FLAGS.problem_type
+    print("Reading %s start" % fpath)
+
+    test_fold = json.load(open(fpath + "folds/test_fold_setting" + str(setting_no)+".txt"))
+    train_folds = json.load(open(fpath + "folds/train_fold_setting" + str(setting_no)+".txt"))
+    
+    return test_fold, train_folds
+
+  def parse_data(self, FLAGS,  with_label=True): 
+    fpath = FLAGS.dataset_path	
+    print("Read %s start" % fpath)
+
+    ligands = json.load(open(fpath+"ligands_can.txt"), object_pairs_hook=OrderedDict)
+    proteins = json.load(open(fpath+"proteins.txt"), object_pairs_hook=OrderedDict)
+
+    Y = pickle.load(open(fpath + "Y","rb"), encoding='latin1') ### TODO: read from raw
+    if FLAGS.is_log:
+        Y = -(np.log10(Y/(math.pow(10,9))))
+
+    XD = []
+    XT = []
+
+    if with_label:
+        for d in ligands.keys():
+            XD.append(label_smiles(ligands[d], self.SMILEN, self.charsmiset))
+
+        for t in proteins.keys():
+            XT.append(label_sequence(proteins[t], self.SEQLEN, self.charseqset))
+    else:
+        for d in ligands.keys():
+            XD.append(one_hot_smiles(ligands[d], self.SMILEN, self.charsmiset))
+
+        for t in proteins.keys():
+            XT.append(one_hot_sequence(proteins[t], self.SEQLEN, self.charseqset))
+  
+    return XD, XT, Y
+
+
+
+class DataSet_for_new(object):
+  def __init__(self, fpath, setting_no, seqlen, smilen, need_shuffle = False):
+    self.SEQLEN = seqlen
+    self.SMILEN = smilen
+    #self.NCLASSES = n_classes
+    self.charseqset = CHARPROTSET
+    self.charseqset_size = CHARPROTLEN
+
+    self.charsmiset = CHARISOSMISET ###HERE CAN BE EDITED
+    self.charsmiset_size = CHARISOSMILEN
+    self.PROBLEMSET = setting_no 
+
+    # read raw file
+    # self._raw = self.read_sets( FLAGS)
+
+    # iteration flags
+    # self._num_data = len(self._raw)
+
+
+  def read_sets(self, FLAGS): ### fpath should be the dataset folder /kiba/ or /davis/
+    fpath = FLAGS.dataset_path
+    setting_no = FLAGS.problem_type
+    print("Reading %s start" % fpath)
+
+    test_fold = json.load(open(fpath + "folds/test_fold_setting" + str(setting_no)+".txt"))
+    train_folds = json.load(open(fpath + "folds/train_fold_setting" + str(setting_no)+".txt"))
+    
+    return test_fold, train_folds
+
+  def parse_data(self, FLAGS,  with_label=True): 
+    fpath = FLAGS.dataset_path	
+    print("Read %s start" % fpath)
+
+    def load_file(file):
+      with open(file, 'r') as f:
+        data = f.readlines()
+        data = [i.strip() for i in data]
+      return data
+    
+    
+    # ligands = json.load(open(fpath+"ligands_can.txt"), object_pairs_hook=OrderedDict)
+    # proteins = json.load(open(fpath+"proteins.txt"), object_pairs_hook=OrderedDict)
+    ligands = load_file(fpath["ligand"])
+    proteins = load_file(fpath["protein"])
+
+    # Y = pickle.load(open(fpath + "Y","rb"), encoding='latin1') ### TODO: read from raw
+    # if FLAGS.is_log:
+    #     Y = -(np.log10(Y/(math.pow(10,9))))
+    
+    # 修改数据输入
+    Y = load_file(fpath["y"])
+    XD = []
+    XT = []
+
+    if with_label:
+        # for d in ligands.keys():
+        for d in ligands:
+            XD.append(label_smiles(d, self.SMILEN, self.charsmiset))
+
+        for t in proteins:
+            XT.append(label_sequence(t, self.SEQLEN, self.charseqset))
+    # else:
+    #     for d in ligands.keys():
+    #         XD.append(one_hot_smiles(ligands[d], self.SMILEN, self.charsmiset))
+
+    #     for t in proteins.keys():
+    #         XT.append(one_hot_sequence(proteins[t], self.SEQLEN, self.charseqset))
+  
+    return XD, XT, Y
+
+

baselines/DeepDTA/source/emetrics.py

+import numpy as np
+
+
+def get_aupr(Y, P):
+    if hasattr(Y, 'A'): Y = Y.A
+    if hasattr(P, 'A'): P = P.A
+    Y = np.where(Y>0, 1, 0)
+    Y = Y.ravel()
+    P = P.ravel()
+    f = open("temp.txt", 'w')
+    for i in range(Y.shape[0]):
+        f.write("%f %d\n" %(P[i], Y[i]))
+    f.close()
+    f = open("foo.txt", 'w')
+    subprocess.call(["java", "-jar", "auc.jar", "temp.txt", "list"], stdout=f)
+    f.close()
+    f = open("foo.txt")
+    lines = f.readlines()
+    aucpr = float(lines[-2].split()[-1])
+    f.close()
+    return aucpr
+
+
+
+def get_cindex(Y, P):
+    summ = 0
+    pair = 0
+    
+    for i in range(1, len(Y)):
+        for j in range(0, i):
+            if i is not j:
+                if(Y[i] > Y[j]):
+                    pair +=1
+                    summ +=  1* (P[i] > P[j]) + 0.5 * (P[i] == P[j])
+        
+            
+    if pair is not 0:
+        return summ/pair
+    else:
+        return 0
+
+
+def r_squared_error(y_obs,y_pred):
+    y_obs = np.array(y_obs)
+    y_pred = np.array(y_pred)
+    y_obs_mean = [np.mean(y_obs) for y in y_obs]
+    y_pred_mean = [np.mean(y_pred) for y in y_pred]
+
+    mult = sum((y_pred - y_pred_mean) * (y_obs - y_obs_mean))
+    mult = mult * mult
+
+    y_obs_sq = sum((y_obs - y_obs_mean)*(y_obs - y_obs_mean))
+    y_pred_sq = sum((y_pred - y_pred_mean) * (y_pred - y_pred_mean) )
+
+    return mult / float(y_obs_sq * y_pred_sq)
+
+
+def get_k(y_obs,y_pred):
+    y_obs = np.array(y_obs)
+    y_pred = np.array(y_pred)
+
+    return sum(y_obs*y_pred) / float(sum(y_pred*y_pred))
+
+
+def squared_error_zero(y_obs,y_pred):
+    k = get_k(y_obs,y_pred)
+
+    y_obs = np.array(y_obs)
+    y_pred = np.array(y_pred)
+    y_obs_mean = [np.mean(y_obs) for y in y_obs]
+    upp = sum((y_obs - (k*y_pred)) * (y_obs - (k* y_pred)))
+    down= sum((y_obs - y_obs_mean)*(y_obs - y_obs_mean))
+
+    return 1 - (upp / float(down))
+
+
+def get_rm2(ys_orig,ys_line):
+    r2 = r_squared_error(ys_orig, ys_line)
+    r02 = squared_error_zero(ys_orig, ys_line)
+
+    return r2 * (1 - np.sqrt(np.absolute((r2*r2)-(r02*r02))))
\ No newline at end of file

baselines/DeepDTA/source/go.sh

+python run_experiments.py --num_windows 32 \
+                          --seq_window_lengths 8 12 \
+                          --smi_window_lengths 4 8 \
+                          --batch_size 256 \
+                          --num_epoch 100 \
+                          --max_seq_len 1000 \
+                          --max_smi_len 100 \
+                          --dataset_path 'data/kiba/' \
+                          --problem_type 1 \
+                          --is_log 0 \
+                          --log_dir 'logs/'

baselines/DeepDTA/source/output/ER/eval_results

+RMSE : 1.485262829572667 ; Pearson Correlation Coefficient : 0.17810684496134926
\ No newline at end of file

baselines/DeepDTA/source/output/ER/results.txt

+6.467522
+6.175456
+5.5424833
+5.7581935
+5.435541
+5.429863
+5.536461
+5.5640984
+5.882646
+5.4769373
+5.488556
+5.625363
+5.313736
+5.808272
+5.6363673
+5.990212
+5.051353
+6.1895947
+6.319407
+5.8177853
+5.7808843
+5.7299786
+5.7299786
+6.176168
+5.6491227
+6.741143
+4.996721
+5.6749225
+4.9942102
+5.852823
+5.673327
+5.219913
+5.455124
+5.543934
+5.543934
+6.3291707
+5.3239236
+6.5188212
+5.399699
+6.225211
+7.1990366
+6.022619
+6.317709
+7.2006974
+6.2298374
+7.3442764
+5.3265986
+6.6267376
+7.236839
+6.809347
+5.998489
+5.5369587
+5.222605
+5.4759645
+4.7184076
+4.7446933
+4.7446933
+4.9453917
+4.8579555
+5.688682
+5.3638024
+6.6970015
+6.922022
+5.2373786
+6.1030083
+5.9858603
+6.2201657
+6.695669
+6.3168545
+5.911902
+7.0040517
+5.64722
+6.207646
+6.664865
+5.4331574
+5.383293
+5.3065886
+5.8029103
+4.62703
+5.880394
+6.6966767
+4.939882
+5.5465236
+5.925244
+6.534122
+6.702632
+6.4006934
+5.208909
+4.9819055
+5.177346
+5.282094
+4.9699063
+4.968585
+5.7913265
+4.2424197
+4.4141154
+4.4141154
+4.504127
+3.9476442
+5.917233
+4.9871855
+5.5201497
+5.50072
+6.2361526
+5.6124153
+6.410565
+5.0378847
+6.1442685
+5.4821434
+5.0435605
+5.075498
+4.989577
+5.8059764
+6.0818005
+4.9550357
+5.2234464
+7.094842
+5.3365226
+6.8286414
+4.292056
+4.778266
+6.6305947
+5.652393
+7.5110865
+5.3670945
+5.2828608
+5.45501
+6.0797043
+6.1816125
+6.1614866
+6.103062
+6.0777783
+6.0552096
+6.1117387
+6.0335836
+7.0962687
+6.6434574
+6.930296
+7.367653
+5.395125
+5.035877
+5.3762827
+6.1513405
+6.1658254
+5.8107777
+5.7238894
+5.7267556
+5.7915087
+5.854144
+5.7156963
+6.226131
+5.684789
+5.570505
+6.7114034
+5.4907975
+6.784401
+6.245473
+5.050832
+5.5548
+6.0875235
+6.6564093
+4.6011233
+5.4594274
+6.0246763
+6.407019
+5.9934154
+5.359483
+5.359483
+5.6769543
+5.6910224
+6.1209445
+5.5472794
+6.472957
+6.8773623
+6.288406
+6.348208
+7.244477
+7.1262827
+6.449307
+6.6500163
+6.635145
+5.7422414
+5.5459576
+6.482847
+6.79241
+7.0690594
+5.228507
+5.5705495
+7.0427613
+5.228507
+5.767734
+6.946606
+5.1587777
+6.2449603
+5.3617554
+7.282804
+5.799242
+6.132366
+6.6706414
+7.5444317
+5.9720287
+6.4760647
+6.1630926
+7.4778113
+7.1809187
+6.490697
+7.2526608
+6.5888877
+6.2422147
+7.1239395
+6.377097
+5.5686545
+6.1431546
+5.850873
+6.8660583
+6.557059
+6.366646
+7.0568414
+6.088338
+7.244094
+5.8533297
+6.2203307
+7.5047336
+5.7495475
+6.666377
+6.3046403
+4.3621182
+5.652291
+8.111586
+6.066658
+6.2466297
+7.560285
+6.1744814
+6.593776
+6.09479
+5.6612544
+5.939238
+7.0114264
+6.1705194
+5.53363
+7.4440317
+6.313761
+5.314995
+6.119054
+6.1605816
+5.9663806
+7.089019
+6.0204525
+6.3210897
+7.2732425
+6.473473
+6.3467903
+6.798541
+6.2466297
+5.4906287
+5.604602
+6.0458856
+6.3662863
+5.373706
+5.722967
+5.6015425
+5.9969525
+6.857416
+5.22418
+5.475809
+6.41353
+6.5152855
+5.2457695
+6.7597604
+5.3962045
+6.0832744
+5.25008
+6.612254
+5.2253714
+6.063141
+5.4104376
+5.572791
+6.1929307
+6.3741198
+6.7629385
+7.553162
+7.4631696
+6.5738697
+6.4958653
+5.3231397
+5.4144382
+6.32794
+7.419338
+7.4119635
+6.5656724
+6.2679086
+6.078931
+6.869893
+6.956627
+6.1463575
+6.0671277
+6.9165044
+5.706223
+6.5210114
+6.0629053
+5.8247347
+6.540231
+7.2961597
+6.316755
+6.149106
+5.897488
+7.3261623
+6.429786
+5.8642225
+7.4559665
+6.094894
+6.3332133
+6.0173855
+6.1059537
+6.1190944
+5.7668066
+5.9509315
+6.3232346
+7.4643135
+7.394734
+7.3374534
+5.5829897
+5.4383006
+6.1083565
+6.768558
+6.7527885
+4.825678
+6.5371857
+5.546511
+5.6412683
+6.221419
+5.4170027
+6.107479
+6.7604346
+5.6207657
+5.266999
+6.107479
+5.537837
+6.997796
+5.4641423
+6.148657
+4.6983767
+5.8901777
+5.946473
+6.5990458
+7.178522
+6.350192
+6.383547
+6.664338
+6.483536
+6.238896
+6.2137227
+6.8680735
+6.239347
+6.8917856
+7.1500893
+5.336475
+6.185486
+5.8683696
+7.3107433
+6.3226795
+5.5120544
+5.8082886
+6.6652327
+6.453342
+7.2482004
+5.4125047
+7.7132316
+5.2179446
+6.0671873
+7.1874647
+6.0671873
+6.366949
+5.462262
+7.044528
+5.051142
+5.9070535
+6.2963076
+6.3234015
+7.3026643
+6.5437765
+5.8900414
+7.804816
+7.0186996
+5.9860735
+6.373691
+6.1429577
+6.501974
+6.8024426
+7.838379
+7.1103787
+6.674358
+6.6560836
+5.846259
+6.8628078
+7.0601616
+6.767179
+5.414812
+6.167569
+5.811791
+6.2401786
+6.136362
+5.629062
+5.39774
+5.7583795
+5.921384
+5.883592
+6.8214016
+6.309103
+7.3759375
+7.180337
+5.689269
+5.732831
+6.8184648
+6.536885
+7.0938787
+5.6343565
+5.703103
+7.208612
+6.1457124
+7.0331993
+5.978574
+6.5694127
+5.490088
+6.3470507
+7.191054
+4.9275794
+6.189489
+7.3133535
+5.484551
+6.9921017
+6.465222
+7.1819453
+7.2947345
+6.1080327
+6.088584
+7.2646194
+7.166979
+6.284159
+5.817567
+5.276581
+5.486908
+5.693727
+6.1640515
+6.6251516
+5.727744
+5.416168
+6.245607
+6.894817
+7.4190426
+7.395757
+5.549125
+6.1100225
+7.7290735
+6.2098575
+5.459329
+5.776312
+5.5599875
+6.7275357
+5.513232
+6.311738
+6.1481915
+5.357658
+5.6171327
+5.504918
+6.066611
+6.1421914
+6.2803297
+5.6643953
+6.0860615
+6.061245
+6.2403326
+5.354926
+6.075662
+5.520243
+6.2785616
+4.694001
+6.130923
+5.403153
+4.9646115
+5.2796474
+5.330424
+5.3950186
+5.4156322
+5.635037
+7.7476335
+4.9557385
+6.0133843
+4.9308424
+5.2614126
+5.4945397
+6.340603
+5.0584426
+5.7315745
+5.065018
+4.2212844
+6.0483475
+5.234022
+4.76302
+5.5551796
+5.23341
+5.210362
+5.110138
+5.5016603
+6.439883
+6.080308
+5.6813083
+6.3253994
+5.819166
+6.5111785
+5.585166
+5.1373324
+5.597088
+6.3559093
+5.7910986
+5.461396
+7.5587125
+5.7162747
+5.302396
+5.75258
+7.333736
+6.084009
+5.0692205
+5.241418
+6.0477796
+5.613452
+6.829801
+5.3483257
+5.9041533
+6.10507
+6.387156
+5.2621484
+6.360398
+5.093281
+6.4135156
+5.3720284
+5.6075544
+5.870746
+6.3124127
+5.5370436
+5.6837006
+5.7438555
+5.536025
+5.910895
+5.8842626
+6.0206275
+4.872495
+6.0826097
+5.8955636
+5.772381
+5.00487
+5.4394894
+5.624068
+7.2013736
+6.2933707
+6.0808144
+5.536216
+5.9835877
+5.1118097
+5.260807
+5.570847
+5.774814
+5.0330424
+5.633846
+5.506184
+4.79763
+5.709769
+5.7201366
+6.673873
+5.5720124
+5.3263655
+5.802265
+5.8956504
+5.971075
+6.1701164
+6.546115
+5.7940593
+5.654686
+5.650679
+6.114338
+6.9369965
+5.005792
+5.2866864
+5.535933
+5.6892242
+6.296921
+5.7082005
+6.0808687
+5.5533624
+5.8792586
+5.7111464
+5.7488804
+5.1158767
+5.456522
+5.2944965
+4.844138
+5.165123
+6.272952
+6.227293
+5.6877546
+6.0242944
+7.09873
+6.280388
+5.0570455
+6.835377
+5.0329375
+6.271915
+6.456241
+6.5826225
+5.275297
+4.9394565
+5.844745
+4.992376
+4.359143
+5.8388567
+6.202779
+5.31049
+5.7588096
+6.6370883
+5.344851
+5.0862412
+4.9653134
+5.4524508
+4.584369
+5.9576993
+5.0613866
+5.6761823
+5.4489317
+5.2452593
+5.9499335
+6.341624
+5.222492
+5.3182554
+6.1915507
+6.6565595
+6.2678857
+6.167898
+5.216836
+4.9330845
+6.163673
+5.761212
+5.777196
+5.4803343
+5.0002365
+5.363882
+4.7241426
+6.001968
+5.011861
+5.2435713
+5.385721
+4.9913692
+5.4485836
+6.195839
+5.2782316
+5.15397
+5.2147355
+3.947084
+5.866563
+5.2386394
+5.3488293
+5.537148
+6.279925
+6.5004563
+6.711674
+5.123827
+4.6733575
+6.192616
+6.230602
+4.963586
+5.543849
+5.2388268
+5.873175
+5.298437
+5.753235
+5.72702
+6.265403
+6.061049
+6.0983605
+5.6412663
+4.365433
+5.0105605
+5.1227226
+5.488341
+5.3545985
+5.026865
+5.5988894
+5.5406065
+4.925748
+5.3008337
+4.4458466
+6.202154
+5.4998484
+5.4405184
+5.9828353
+6.2485967
+5.361987
+6.252941
+5.6222343
+5.592175
+5.91426
+5.3488817
+5.0730023
+6.5664697
+5.6709094
+5.4205346
+5.4755836
+5.9263487
+6.1135435
+5.2365828
+6.389742
+5.8542604
+5.555178
+5.5159826
+6.2687693
+5.486962
+6.400485
+4.9783463
+5.6938205
+6.247785
+5.1759753
+4.9569697
+6.900387
+5.211111
+5.1996584
+5.251429
+5.4088726
+5.475929
+4.3215995
+4.6796637
+5.9881783
+6.951918
+5.158981
+4.9406147
+6.8510613
+4.1983995
+5.2218494
+5.8812246
+5.5123997
+5.194318
+4.9121003
+5.144572
+5.638291
+5.4011235
+5.3436866
+5.499963
+5.5898023
+5.722905
+5.687527
+5.5680947
+5.8099523
+5.5097027
+6.48778
+5.736049
+5.2097087
+5.1820335
+6.6507125
+6.616532
+6.221067
+6.3871517
+6.3871517
+6.286934
+6.9081984
+6.286934
+6.9081984
+6.6990314
+6.320643
+6.2955394
+6.5417485
+6.419246
+6.7017126
+5.972306
+5.593537
+5.43418
+7.1164546
+5.905944
+6.4156313
+5.312497
+5.945516
+6.455089
+6.5600963
+5.989281
+6.6487565
+6.845637
+7.55256
+6.2125487
+6.355933
+6.262196
+6.369975
+6.230882
+6.320643
+6.436056
+6.3782277
+6.570935
+6.489056
+6.5716195
+6.3782277
+6.3010693
+6.5750556
+7.3701096
+7.0298605
+7.1175966
+6.292645
+6.0925097
+6.094389
+6.2471895
+6.1777945
+6.2675815
+6.659286
+5.6245112
+7.3324537
+5.670894
+4.9676332
+5.06271
+4.7458615
+6.5478196
+5.2051744
+5.213387
+5.0545807
+6.2032223
+4.933809
+6.7999773
+6.37943
+6.436173
+5.7264333
+6.7405252
+5.573129
+5.869438
+5.4759645
+6.378834
+6.333488
+6.100217
+5.507454
+5.4479175
+4.9235926
+5.3989277
+5.085065
+5.908596
+5.7883053
+7.2705007
+6.94589
+6.8676815
+5.3989096
+5.220262
+5.0723066
+4.967504
+5.1772046
+4.947744
+5.631586
+5.166439
+5.656475
+5.447201
+5.5681186
+6.291062
+5.803462
+6.7605786
+5.9466567
+5.932242
+6.043377
+6.3112845
+5.3567986
+6.4313383
+6.067965
+4.9574113
+6.5868063
+6.824912
+7.011283
+5.621954
+6.5735836
+6.233931
+5.930412
+6.700928
+5.7168355
+5.627964
+5.7621074
+5.531625
+6.2479553
+5.6877246
+5.961213
+5.1570897
+6.146668
+5.6236563
+6.710904
+6.6558375
+5.3288183
+5.313829
+6.6436534
+5.250808
+5.056483
+6.291795
+5.5476236
+5.373606
+5.647381
+6.8223333
+5.8831334
+5.464456
+5.223276
+5.7664294
+5.005566
+6.543031
+5.7644773
+6.0466204
+5.197456
+5.066023
+6.1397996
+5.0928364
+5.449444
+5.5406613
+5.302817
+5.7155867
+5.690205
+6.085121
+5.6479006
+5.968774
+5.277886
+4.9721637
+5.141983
+5.868268
+5.0960402
+5.4958825
+6.888697
+5.689254
+7.246741
+6.4569216
+7.246741
+6.7176867
+5.96806
+7.246741
+5.127513
+6.4972453
+6.092105
+5.9237747
+6.476934
+7.319737
+6.2266574
+6.9943686
+6.503329
+5.993003
+6.2854657
+5.5388966
+7.033675
+5.250325
+7.0303593
+5.784215
+5.7090406
+5.5398903
+5.4250836
+5.4678726
+5.880919
+4.5848045
+4.5973973
+5.8356805
+4.8457956
+6.2677946
+5.804833
+6.1843605
+5.511601
+5.745906
+5.9973736
+5.3361974
+6.899275
+5.052643
+5.3279963
+6.2756667
+6.748947
+6.4433417
+5.4147453
+5.7602544
+5.995493
+6.517574
+5.7964764
+5.756449
+5.7015796
+5.3037014
+6.013344
+5.8427305
+6.1242275
+5.328395
+5.36347
+4.0370135
+5.7501745
+6.083869
+6.586791
+6.8433857
+6.5928197
+7.375711
+4.184068
+4.9256296
+5.274472
+6.4419727
+5.377065
+5.3647532
+6.4676127
+5.707027
+6.492631
+6.507264
+6.786269
+5.4085755
+5.961538
+6.421152
+6.009842
+6.184998
+6.194706
+5.8600044
+6.194706
+5.766577
+5.506866
+7.148358
+7.73028
+5.235998
+6.7807255
+7.6936836
+7.677403
+6.0699534
+5.7997165
+7.2714953
+5.933518
+5.5884705
+5.967518
+4.85046
+6.3284354
+5.889121
+6.0699534
+6.2526226
+6.8087096
+5.1664944
+6.893789
+7.710523
+6.2668576
+5.4497004
+4.7879705
+6.4280524
+6.633815
+6.5708776
+6.8266087
+5.4514575
+5.875547
+7.6783714
+6.2668576
+5.0025067
+6.1850924
+5.308676
+7.1538615
+6.960084
+7.5788445
+5.6230187
+7.233242
+5.308676
+7.1039195
+5.5000815
+5.891763
+5.494062
+5.9090996
+6.8754225
+7.266729
+5.9655166
+7.469999
+5.796821
+7.469951
+5.4577727
+7.266729
+7.1538615
+5.500397
+5.809769
+5.6330376
+5.90072
+5.5874615
+6.1716623
+6.022232
+5.890116
+5.891763
+5.8908978
+5.830562
+6.005416
+6.7033906
+6.6815243
+5.962167
+6.2592874
+6.335843
+7.4758735
+6.6545043
+5.699082
+5.7970614
+5.948544
+6.4484124
+5.9513273
+6.690463
+5.757504
+6.0261083
+5.9205647
+7.078332
+5.9386554
+5.5954394
+5.913443
+5.7262807
+6.020115
+6.7479596
+7.0793085
+6.143913
+7.337945
+6.498925
+6.609576
+5.9278646
+6.4115324
+6.9145627
+5.4193535
+5.529038
+5.658442
+7.3179717
+6.8085446
+7.2937202
+6.300633
+6.0707784
+6.425224
+5.8391795
+7.0364804
+6.802596
+5.657187
+6.4053726
+5.361449
+5.6154866
+6.710414
+5.7986093
+6.180119
+5.641009
+6.608839
+5.8286757
+6.857249
+5.192629
+6.9102287
+5.601957
+7.016434
+7.333718
+5.3080096
+6.767535
+5.146874
+7.1538615
+7.0937257
+5.817904
+7.1551766
+5.262034
+6.5052304
+5.308676
+6.5945153
+7.3038073
+5.4937625
+5.4190507
+5.19725
+5.4331875
+5.551677
+5.8899326
+5.1104307
+6.786726
+5.603208
+6.4982767
+7.0305605
+6.9125113
+6.8070955
+6.617151
+6.819119
+7.043898
+6.9137826
+5.2608347
+5.3874426
+5.43459
+6.499948
+5.4580803
+5.991781
+5.9642334
+7.216374
+5.8875966
+7.288489
+6.671123
+5.9083867
+7.723083
+7.323798
+5.9325633
+5.7545996
+7.660396
+4.497191
+6.6828775
+6.8184133
+6.565157
+8.335036
+7.280432
+6.1306925
+5.6986794
+6.42279
+7.158138
+5.1008873
+5.947463
+6.8608756
+6.757793
+5.6224194
+7.133646
+7.478069
+7.3342133
+5.6451106
+6.8659325
+7.2923803
+6.1306925
+6.813074
+5.6451693
+6.6828775
+6.11627
+5.6391244
+6.8865876
+5.285064
+7.5761304
+7.8262296
+7.413522
+7.5259423
+6.931205
+5.7975683
+6.931205
+6.427707
+5.7830963
+7.891785
+5.355285
+6.575018
+6.205735
+5.4644556
+7.4474983
+5.81593
+7.4846444
+7.1167746
+5.355285
+5.4839234
+6.5376797
+6.507361
+5.8048496
+7.158224
+7.40506
+7.189181
+6.0103397
+5.3693852
+6.033007
+7.411433
+6.7043843
+5.350382
+6.0198426
+7.313643
+5.242876
+6.7147474
+6.1955814
+6.201557
+6.033007
+6.895689
+6.94712
+5.5774407
+6.552561
+6.1911273
+5.3538375
+6.4487123
+6.7837596
+6.6780787
+6.46029
+5.9350615
+6.3753114
+6.220034
+5.4951243
+6.2052674
+6.199719
+5.350382
+7.470936
+7.1710763
+5.355625
+7.1031995
+7.4039483
+5.5698967
+5.886789
+5.8479943
+6.0936513
+6.4315224
+6.5594063
+6.4986777
+6.689513
+5.0134554
+6.1502748
+5.747095
+6.163845
+5.749958
+6.376411
+6.679356
+7.57837
+6.3566575
+5.8474493
+6.0372753
+6.6605186
+6.0885553
+6.027188
+6.6392574
+6.665908
+6.939761
+5.742042
+7.1039853
+6.943179
+6.93992
+6.183739
+6.2344203
+5.494151
+6.3465204
+5.583018
+6.922671
+6.0705886
+5.8718204
+6.9564886
+6.3693786
+7.1693697
+6.714978
+6.911027
+6.6053348
+6.863506
+5.808535
+5.363819
+6.82491
+7.2623053
+5.7913437
+6.70864
+6.207999
+6.9364448
+6.6619096
+6.8284225
+6.6435046
+6.0850997
+6.819591
+6.1630807
+6.553617
+6.414538
+6.4792995
+6.830395
+6.1988306
+6.70383
+5.893571
+7.480984
+6.038288
+6.088005
+6.225726
+6.32311
+6.6287565
+7.309803
+7.3003874
+6.6930227
+6.7055793
+6.5289474
+6.9721994
+5.653107
+7.168499
+5.2066326
+7.397833
+7.5011764
+7.180694
+6.367112
+6.9869504
+7.1919727
+6.5436597
+6.739212
+7.4145775
+6.581346
+6.890866
+5.590059
+6.701334
+6.2036176
+7.018711
+7.4155045
+6.621374
+5.4439697
+6.0163956
+6.1863523
+6.7419724
+6.14213
+5.600849
+6.889785
+6.900848
+6.8721566
+6.1751094
+6.1567
+7.4597073
+6.334549
+7.5797625
+6.615716
+7.4721603
+6.3857384
+5.6308236
+7.3923736
+6.861946
+6.1686893
+6.5457377
+6.863506
+6.708418
+7.3462076
+6.4951878
+6.6720157
+6.621374
+7.3204875
+5.881873
+7.1606784
+6.4986777
+6.5406203
+6.802854
+6.980467
+6.2821507
+6.9008846
+6.849551
+6.9547825
+6.101282
+6.113383
+6.761018
+5.404313
+7.764529
+5.399896
+5.807212
+5.5574293
+6.936211
+5.5323634
+7.32933
+6.9930487
+6.8210154
+6.7600226
+6.761018
+5.5657625
+5.404313
+5.5657625
+7.4274545
+5.880738
+5.8185554
+6.936211
+6.7175026
+5.6989126
+6.97765
+5.5470405
+6.8210154
+5.8185554
+6.97765
+5.5574293
+7.764529
+5.6989126
+5.78929
+5.5657625
+7.2127204
+5.880738
+6.9452853
+7.734715
+5.516134
+6.7001934
+6.5111785
+6.9696507
+7.279464
+6.8160696
+7.0648828
+6.47734
+4.892709
+6.838542
+5.8903117
+5.8013043
+6.8162866
+6.095815
+5.2835183
+6.1388106
+6.109304
+5.336598
+5.2365756
+5.575399
+5.9057035
+5.8066893
+5.9752703
+5.636675
+6.0353594
+5.3802185
+5.6251574
+5.6704345
+6.209346
+5.9591203
+5.8158846
+5.7781544
+6.062674
+5.98491
+5.931804
+5.9149375
+5.553958
+6.720748
+5.8682327
+5.7852535
+6.119146
+5.6444488
+5.579783
+5.8886986
+5.8848295
+5.778422
+6.050983
+5.541333
+6.1166854
+5.883996
+5.420611
+6.6455793
+6.6455793
+5.8289485
+6.1575556
+5.9673476
+6.44633
+5.8289485
+5.470754
+5.822579
+6.4728546
+6.834802
+6.7331834
+6.486193
+7.068154
+6.87085
+5.0722394
+6.44633
+6.897445
+6.238226
+7.011872
+5.0722394
+6.452281
+5.420611
+5.7927575
+6.795473
+5.210459
+6.2383
+6.795473
+5.210459
+5.0722394
+6.897445
+6.3392963
+5.9988317
+6.9227586
+6.238226
+5.8289485
+6.9227586
+6.491039
+6.413065
+6.5267844
+5.6489882
+6.5463734
+6.221266
+6.4356327
+6.492518
+5.191523
+5.5414495
+5.0106807
+5.671316
+6.2691092
+5.3167467
+4.599587
+6.686434
+6.904624
+5.8677015
+5.136459
+5.28131
+5.5673103
+6.093651
+5.586382
+5.614611
+6.299219
+6.267537
+6.4146814
+5.166318
+5.6953497
+5.1839194
+6.616153
+5.7625847
+6.449054
+6.4484005
+5.2185993
+5.0258117
+6.510044
+6.2456594
+6.3107443
+5.253698
+4.874043
+7.0243454
+6.813769
+5.0963154
+6.1877913
+6.256463
+5.983534
+5.1863794
+5.6602716
+5.964029
+5.453381
+6.30822
+5.841526
+6.6644497
+4.994491
+5.104412
+5.8758774
+6.279933
+5.284153
+4.6679087
+6.3260183
+6.086885
+5.315095
+5.96516
+6.289207
+5.3958697
+6.084689
+8.099635
+4.9490347
+6.0606585
+6.989985
+6.996923
+6.609996
+6.6182337
+6.9544344
+6.7357354
+5.8761187
+6.306437
+5.1342115
+7.0245657
+7.105329
+5.714878
+7.656184
+6.3806853
+5.37554
+5.5982084
+6.7982135
+5.150727
+6.161255
+6.629156
+5.7284007
+6.578259
+6.8417006
+6.838354
+6.675882
+4.937349
+5.436128
+7.474793
+7.087658
+6.6610703
+5.464772
+6.625414
+6.1665034
+6.930605
+5.976609
+6.8952274
+5.7105894
+6.0348577
+5.8940177
+5.6628723
+6.7803946
+5.2164364
+6.085289
+5.243691
+7.301918
+5.984678
+6.246605
+6.6402817
+6.6718707
+5.058132
+6.271823
+6.5274215
+7.1961713
+6.4966507
+6.600976
+6.814276
+5.9806075
+5.921116
+7.263263
+5.3209357
+5.8854475
+6.9353766
+5.8740463
+6.7519217
+6.0534625
+7.3050323
+4.8728805
+6.2462926
+6.193253
+6.9410944
+7.206945
+6.8756843
+4.8776455
+7.161649
+6.684436
+6.1178584
+4.964346
+6.058733
+4.9377127
+7.071442
+6.4998837
+5.9685774
+6.2168183
+6.122586
+6.344562
+6.93882
+6.207654
+6.8316703
+6.2324185
+7.4262877
+6.852638
+6.9488473
+6.967378
+5.7041483
+5.752214
+6.9887104
+5.8168845
+5.366215
+6.4164557
+6.99331
+6.3117337
+5.912122
+6.034814
+5.4360113
+6.3156056
+5.768529
+5.2735353
+5.8346186
+6.3373537
+6.710958
+6.3504143
+6.7891693
+5.733319
+7.1769357
+6.5388384
+5.989903
+5.125636
+6.233138
+6.3925943
+5.968524
+6.3226624
+6.5813217
+6.171594
+6.1498837
+6.390574
+5.9392643
+5.3613753
+5.3613753
+6.6410275
+5.3675675
+5.3613753
+6.634193
+6.1924634
+6.391904
+6.871654
+6.219294
+6.629328
+6.896392
+7.01068
+6.6230016
+6.4824905
+6.411081
+6.737173
+7.0029035
+6.5056505
+7.084986
+6.6073723
+6.4475627
+6.8394356
+6.7053137
+6.9089494
+6.9577923
+8.258459
+6.417294
+6.888794
+7.0931168
+5.9251447
+6.390502
+5.8794794
+7.0603194
+7.023944
+6.0059996
+6.430966
+6.293146
+7.1824613
+6.755937
+5.7323427
+6.4565535
+6.799849
+6.4493523
+7.002536
+6.745455
+6.3426175
+6.112106
+5.8313427
+6.957606
+6.9475193
+6.5767937
+6.4443398
+6.692452
+6.1842217
+5.4548492
+6.7909026
+7.249493
+6.330179
+6.2062163
+6.2923346
+6.204242
+5.9851522
+5.2534046
+6.065061
+5.967025
+4.7989936
+5.1041102
+5.1041102
+5.703567
+4.978741
+6.6864715
+5.451647
+6.2062163
+5.1269336
+5.5204434
+5.869374
+5.483179
+6.054952
+6.0446043
+5.8325305
+6.161826
+5.9581776
+6.032217
+5.656331
+5.1506643
+5.3290267
+6.3243265
+5.403124
+6.7040186
+5.7676363
+5.347377
+5.5251346
+6.179441
+6.282246
+5.9100685
+5.9557357
+6.6209874
+5.7568254
+5.226681
+5.086042
+5.084005
+6.8140683
+5.8810186
+6.013246
+5.8042626
+6.469317
+6.7973375
+4.9946556
+5.5528173
+5.587806
+4.7347746
+5.4313626
+5.963972
+5.55007
+4.9913373
+6.499063
+6.198709
+5.7344494
+7.164914
+4.91922
+5.962316
+5.594864
+4.726662
+5.3363442
+5.303285
+5.0192604
+4.9645977
+5.559672
+5.0866556
+5.702638
+5.6967287
+4.817513
+5.1762395
+6.8561244
+6.0643682
+5.52713
+5.0605087
+4.628874
+6.078061
+5.3982654
+5.8325305
+6.725618
+5.285354
+6.594067
+5.810192
+6.011201
+6.21154
+5.524951
+4.939838
+4.860134
+5.4144764
+5.152231
+5.2471013
+5.5232177
+6.279133
+5.612306
+5.8928256
+4.715712
+6.3039327
+5.1076765
+6.1576266
+6.299791
+4.497044
+6.1354585
+5.3971553
+4.7871237
+5.837885
+5.6793804
+6.369365
+5.8135138
+6.672243
+5.2631044
+5.333174
+5.70223
+5.0657125
+5.8304696
+6.381367
+5.7556
+5.70386
+5.784319
+5.225506
+5.737361
+5.878706
+5.2299895
+5.5232844
+4.896603
+5.8859997
+6.8943076
+6.884344
+7.0199056
+6.8824687
+6.168615
+5.6734533
+5.938505
+5.8402905
+7.5200944
+6.2119412
+5.9067955
+5.6050105
+5.2653317
+5.2059035
+5.3764734
+7.3070383
+5.121628
+5.302547
+5.1488953
+5.073028
+6.330474
+5.5348835
+6.254263
+7.032158
+6.266541
+4.8056045
+6.185329
+4.8056045
+6.210166
+7.032158
+6.5320005
+6.1325426
+5.3772693
+7.032158
+6.045276
+5.4125586
+6.1912265
+6.5712376
+6.696406
+6.5712376
+5.881195
+6.696406
+7.1256957
+6.266541
+5.2304263
+5.968301
+5.5359416
+5.946769
+6.4270353
+5.306702
+6.241101
+6.5984855
+6.4078135
+6.0889125
+6.4195113
+6.5750647
+6.564784
+5.58414
+5.5847282
+5.7705774
+5.9008756
+6.248256
+6.449347
+6.696406
+6.302776
+6.5434055
+4.715931
+5.7887897
+5.3597546
+5.9008756
+5.3460603
+6.629929
+7.207822
+5.8611436
+6.5493627
+5.1872463
+7.207822
+6.5493627
+6.206546
+6.6267595
+6.3123145
+7.5563545
+6.745877
+6.730434
+6.305086
+6.2199597
+5.887241
+6.093286
+6.54374
+6.803407
+6.1036186
+6.372928
+6.99878
+6.2782784
+5.7770004
+6.8059063
+6.104658
+5.867501
+6.714566
+6.619787
+5.639694
+6.827391
+5.0718384
+6.7611136
+6.4251614
+6.0556
+6.5472074
+6.7874436
+6.392543
+6.572409
+6.093286
+6.730434
+6.7283907
+6.7611136
+7.020267
+6.421667
+6.305086
+6.8642035
+6.709178
+6.4529314
+5.9857283
+5.9857283
+5.9857283
+5.9857283
+5.9857283
+5.9857283
+7.633851
+7.3244433
+6.7036314
+6.310724
+6.5329785
+6.5929947
+5.8618703
+5.706366
+6.3075547
+7.196339
+5.8618703
+7.4691744
+6.266541
+7.196339
+7.1543555
+4.715931
+7.2166243
+6.449347
+4.8056045
+5.705721
+7.1543555
+5.9802284
+7.4237223
+5.705721
+5.909997
+5.8047705
+6.408692
+6.675869
+6.8563094
+6.2897015
+6.208653
+4.9841747
+7.753767
+7.2476044
+6.3989143
+6.8494897
+5.382583
+5.2271442
+5.6690764
+5.850965
+6.150583
+5.4191294
+5.4166822
+5.9547234
+5.5940204
+5.12253
+5.8302155
+6.1483936
+6.6816854
+5.729907
+6.02498
+6.190885
+5.7683287
+5.8202615
+4.7404923
+5.6471686
+4.830724
+5.735604
+6.503758
+5.480142
+5.4643526
+5.913085
+6.2220254
+6.46555
+5.8918805
+5.812311
+5.621966
+5.440773
+5.9304795
+5.9624453
+5.778366
+5.5331364
+6.6829753
+5.2990212
+6.7413177
+5.9778285
+6.110689
+6.6678534
+6.0362663
+6.3602934
+7.2238183
+6.1834545
+6.09217
+6.6472893
+7.2145658
+6.609166
+6.4584284
+5.7264094
+6.004381
+6.3551064
+6.86889
+7.275844
+6.0574303
+6.273188
+7.361916
+6.518585
+5.7264094
+6.089704
+6.945603
+6.7924833
+5.9155984
+6.8057795
+6.273188
+6.996869
+6.3595424
+5.7264094
+6.937701
+7.157386
+7.0525227
+7.176013
+6.3581676
+6.7461414
+6.030363
+6.273188
+7.4644194
+6.1666574
+6.4856677
+6.824707
+6.0735745
+4.8937964
+7.4275713
+5.842707
+5.590454
+6.665908
+6.1751094
+5.4439697
+5.8474493
+5.2686744
+5.3236814
+6.2928267
+5.6969624
+7.7352457
+6.8126435
+6.5901375
+5.1797843
+6.3946877
+6.064659
+5.639618
+5.5225515
+6.3346405
+6.4140244
+5.7628813
+5.7628813
+7.073425
+7.1590343
+6.4880567
+6.5973787
+6.4605
+6.5495114
+6.4723654
+7.4309497
+6.4920187
+5.363725
+5.363725
+6.545209
+6.783951
+5.9570446
+6.2748013
+6.2770853
+6.3901615
+6.2841473
+6.705341
+6.2602153
+5.9027004
+5.599551
+5.9539185
+6.516948
+5.836895
+5.720774
+6.421318
+6.634815
+7.008399
+7.3101745
+7.436628
+5.791473
+6.443822
+7.540551
+7.6198635
+6.3792706
+6.577402
+6.0336556
+6.6158924
+6.7791862
+5.207016
+6.7588897
+7.844326
+6.2823796
+6.703623
+6.3063498
+7.3694186
+6.7811313
+6.801373
+6.9923296
+6.7802753
+6.415253
+6.406323
+7.19898
+8.035133
+7.239584
+7.2319098
+6.855592
+7.358061
+7.7228737
+7.3471255
+6.6850677
+5.441019
+6.7229156
+6.3937416
+7.397027
+7.01828
+5.386232
+7.442019
+6.870635
+5.9424877
+5.959657
+6.917614
+6.3962884
+5.608596
+7.933974
+7.8642387
+7.593652
+6.7859898
+7.6548233
+7.0926714
+7.4650073
+7.274835
+6.8358603
+7.576113
+7.26126
+5.768214
+6.202369
+6.0876827
+6.538478
+5.649271
+5.921262
+6.3254766
+6.147927
+6.515933
+5.9064384
+6.538478
+6.846441
+5.921262
+7.2793765
+7.322479
+6.0876827
+6.846441
+6.16077
+7.322479
+6.829197
+6.874249
+5.9064384
+6.2150803
+7.0031123
+7.142586
+6.4135585
+5.916401
+5.3300714
+6.16077
+7.0092688
+6.3254766
+6.041011
+6.147927
+5.160581
+5.649271
+7.0092688
+6.9832053
+7.0031123
+7.142586
+6.2150803
+5.916401
+6.515933
+6.874249
+5.776429
+5.179457
+5.160581
+5.3300714
+6.147927
+5.179457
+6.4135585
+6.9832053
+6.5818276
+5.776429
+7.65558
+5.1435676
+6.1266737
+4.9849977
+6.0107045
+6.124716
+4.896845
+7.310674
+6.701244
+5.5520854
+5.8173804
+6.836
+6.508478
+5.8555493
+6.7545896
+6.061616
+5.5693445
+6.366635
+5.9226646
+6.3179035
+5.6914854
+7.0459146
+5.9947195
+5.5198374
+6.7556376
+5.488785
+6.6878595
+7.0701385
+5.632714
+5.488785
+6.6878595
+5.479324
+4.8022003
+6.549372
+5.4726763
+4.796402
+5.2994943
+5.4624367
+5.1461525
+5.3280177
+7.1630106
+7.189575
+7.323072
+5.6898046
+5.764552
+6.2346563
+5.0439134
+5.566765
+5.8321614
+7.482371
+5.654193
+5.0227966
+5.4175177
+5.793097
+5.1891603
+7.249434
+6.9494905
+6.692452
+6.7909026
+5.8313427
+6.4443398
+5.117519
+4.969426
+6.3044095
+7.1654277
+6.978273
+5.7630506
+6.1041512
+5.816536
+7.1312404
+6.3851147
+7.3398256
+5.9146333
+5.793097
+5.5414186
+5.5925574
+7.1681724
+5.5963306
+6.2440205
+5.9414506
+6.2716227
+5.342479
+6.002595
+4.8309937
+6.3972225
+6.441071
+6.5236244
+6.4115953
+6.481668
+5.4318457
+6.335093
+5.051272
+6.789253
+4.9637017
+5.997874
+5.3621125
+6.7521963
+5.549838
+5.2304025
+6.458685
+5.4439697
+4.5430627
+5.1819153
+4.936701
+6.6406226
+6.013207
+7.1670628
+5.791503
+5.8600044
+6.5929947
+5.0783725
+5.7297564
+7.151771
+5.792663
+6.3942857
+5.7756777
+5.3302484
+5.3856387
+4.998325
+6.3902006
+6.4683266
+5.9130745
+5.697284
+6.3861585
+5.1710477
+5.84126
+5.054979
+5.161748
+6.8855643
+6.509545
+4.9780416
+4.6288743
+5.893382
+5.458822
+6.5929947
+5.396231
+6.0478344
+7.151388
+5.472148
+5.3613753
+6.0795956
+5.954827
+5.6403384
+6.8229175
+5.96385
+6.2803507
+5.4260345
+6.003474
+5.90873
+5.068148
+6.7081003
+5.6237574
+5.9156036
+6.20552
+6.1944494
+5.103165
+5.448076
+6.229413
+5.321262
+5.42142
+5.191296
+5.5573125
+6.183935
+6.9862604
+5.9582157
+6.7052107
+5.578901
+5.5467825
+5.4386425
+4.657817
+5.4645243
+5.394047
+5.680804
+5.23331
+5.632147
+5.3293037
+5.6983004
+5.2864676
+5.458678
+6.553249
+5.8577437
+5.387469
+5.1316023
+5.258217
+6.5670776
+5.4777985
+6.364756
+5.8544817
+5.3393674
+6.5614405
+5.806353
+6.805071
+6.8549705
+6.6222363
+5.329012
+5.3986382
+5.580193
+5.5056796
+5.126788
+6.4482384
+5.718361
+5.067653
+5.5428834
+5.4606895
+4.830859
+5.467882
+6.005772
+5.458219
+6.384924
+4.9810133
+5.417175
+5.5332274
+5.594346
+5.6218033
+4.785696
+5.573224
+5.3005395
+6.0044074
+5.3885126
+5.9977345
+4.6011233
+5.969474
+6.3812084
+6.3211975
+6.5309973
+6.7326913
+4.94913
+5.924118
+5.5111303
+5.6410227
+5.7037992
+5.687299
+5.4345837
+5.507501
+5.803969
+5.4489136
+5.8247304
+5.4088726
+6.0947556
+5.258217
+5.3665876
+6.5929947
+6.489723
+5.112482
+6.3868294
+6.331831
+6.356875
+5.7077713
+6.6891823
+5.7072167
+4.5430627
+5.393888
+5.1954904
+6.254587
+6.9102287
+5.5712314
+5.3885126
+5.7756777
+6.545688
+5.4136405
+4.9245124
+5.3926406
+6.9398413
+5.4600725
+5.2927065
+5.8954806
+5.758978
+5.1536565
+5.906547
+5.1954904
+6.1635704
+5.613791
+6.3214912
+6.2841077
+5.404313
+5.649161
+6.6616483
+5.716136
+5.8577437
+7.386524
+5.3898168
+5.8805575
+5.4844666
+6.696406
+4.296452
+7.231281
+6.3793616
+5.738969
+6.78125
+5.643633
+6.3860025
+5.9562473
+4.8628693
+5.2466655
+5.6291747
+6.32467
+5.0683675
+6.517736
+5.955325
+6.654591
+6.2360516
+4.8251915
+4.9508615
+5.8747706
+5.575154
+6.179828
+5.515387
+5.732545
+4.9928393
+5.7124543
+4.9841185
+5.6373434
+5.210205
+6.7779403
+4.785696
+5.2912946
+6.337707
+5.6388702
+4.8755174
+5.697323
+4.913454
+5.872951
+5.143903
+6.3294034
+4.2287254
+5.9581776
+5.500564
+5.4695253
+6.1849694
+6.145796
+5.5224404
+4.9601436
+6.7825117
+6.2582684
+5.171891
+5.4694295
+5.7206063
+5.7361574
+5.7455497
+7.101007
+5.4775743
+5.8247304
+5.967657
+7.275006
+6.2976384
+5.5106735
+6.616783
+5.792663
+5.2035565
+5.3227415
+5.8785
+5.7449207
+5.520708
+5.810928
+4.972129
+4.9813147
+6.3966904
+5.482785
+5.63081
+5.8641477
+6.6450086
+5.352088
+5.3322296
+5.165727
+4.5430627
+5.579981
+6.1755667
+4.89529
+6.4848285
+5.1645045
+6.1189384
+6.1077356
+6.1308722
+5.7454777
+5.270401
+6.111031
+5.404313
+7.406449
+5.5731516
+5.0669465
+6.111031
+5.2109985
+5.662803
+6.9715033
+5.970828
+5.539181
+5.7307296
+5.6358924
+6.537821
+5.365324
+5.883437
+5.8760376
+5.3372855
+5.9737754
+7.143226
+5.515236
+6.6628046
+6.660112
+5.5342803
+4.8639097
+5.8131795
+5.2035565
+5.7094584
+5.962167
+5.0812287
+5.0690403
+5.991651
+6.050492
+5.357571
+5.1913004
+5.3590655
+5.901271
+6.1262636
+5.563857
+6.5929947
+6.949256
+6.1830287
+5.3565626
+6.7407084
+6.5436597
+4.9169283
+6.649489
+4.8056045
+6.7903647
+5.291184
+6.771679
+5.190703
+7.2657433
+5.2911468
+5.315289
+5.027199
+6.209844
+5.120556
+5.4544845
+5.2825212
+5.467882
+4.9815717
+6.0703673
+5.2565007
+5.616389
+4.7857103
+4.9852576
+6.829197
+5.352088
+6.1400914
+5.5977283
+5.857426
+5.5487347
+5.187174
+5.2864676
+5.3120327
+5.1604357
+5.987527
+5.55243
+6.7486444
+5.6913486
+6.660112
+6.034846
+5.304135
+5.4954214
+5.7780843
+5.4354787
+5.7176723
+5.578759
+5.351402
+5.594662
+5.468805
+4.7305474
+5.497133
+6.227457
+4.7853174
+5.4288287
+6.1657615
+5.8360243
+5.198638
+6.1765084
+6.189051
+5.415945
+6.155344
+4.9300656
+5.583409
+5.9787865
+5.614335
+5.258217
+6.449
+5.7538176
+5.401879
+4.8056045
+5.208134
+4.861804
+5.158764
+5.520353
+5.6117415
+5.728972
+5.7927938
+5.739141
+7.857526
+5.454875
+5.8247304
+5.4733872
+5.2669387
+5.1728573
+5.7257504
+4.7857103
+5.017933
+5.176077
+5.4600725
+4.5430627
+5.509913
+7.0282617
+5.624545
+6.549943
+5.0664043
+5.0360723
+4.927418
+6.964592
+6.0670524
+5.357271
+4.973726
+5.8764157
+5.8323
+5.191296
+5.455521
+5.6581197
+5.5000715
+5.6792316
+5.233205
+5.817236
+5.6474333
+5.405328
+6.1796565
+5.467526
+5.3885126
+4.900526
+6.522156
+5.208627
+5.1417003
+5.11362
+6.704824
+5.90873
+5.886378
+5.112907
+5.90873
+5.305006
+5.221336
+5.6202893
+4.8586655
+5.525492
+5.952324
+5.7454777
+5.411204
+7.1086564
+5.3362865
+6.007184
+5.7276807
+5.317792
+5.200025
+7.087147
+6.5142217
+5.382159
+5.4394507
+5.4737973
+6.3902006
+5.1400824
+5.3565626
+6.0822306
+5.924946
+5.1305737
+5.7658415
+5.2723975
+6.265974
+6.456534
+6.6053085
+6.718813
+5.3699093
+5.69949
+7.090281
+6.0329123
+6.279035
+5.1232476
+5.9395356
+5.258217
+5.2768917
+7.0589504
+6.0967617
+5.940691
+5.9540744
+5.1727777
+6.2118473
+5.1320553
+5.7810836
+5.179801
+5.485778
+6.308994
+5.655888
+6.1280494
+5.6262813
+5.6282444
+5.3190145
+5.8162494
+5.593325
+5.198212
+5.791503
+5.8796315
+5.282864
+5.2723975
+5.5311418
+5.501936
+5.526906
+5.7457466
+6.0266857
+5.613171
+5.305006
+6.3531313
+5.054123
+4.923471
+5.4268613
+5.906971
+5.2669387
+6.4360414
+5.80696
+6.133394
+5.2304025
+6.7693343
+6.303871
+7.190364
+7.551679
+6.099068
+6.5598974
+6.200362
+7.1434317
+6.893181
+7.6898084
+6.8266087
+6.166194
+6.8279295
+6.0852942
+5.987242
+7.0758677
+6.3492208
+6.117755
+6.176795
+7.226082
+6.406439
+7.6452026
+5.6524563
+6.0194936
+7.056963
+6.997655
+7.356374
+5.8845797
+6.166194
+5.5555096
+6.606661
+6.8870244
+6.941169
+5.3704224
+6.986261
+6.3947344
+6.9378343
+6.7914295
+7.489514
+6.70887
+7.3117642
+7.8877735
+6.663147
+6.04455
+6.4844275
+5.6994276
+6.239663
+6.3301196
+6.73893
+5.7858152
+7.7444754
+6.958664
+6.5268927
+6.113218
+7.143577
+6.8727345
+6.1731687
+6.7365127
+6.9751334
+6.9013796
+6.777693
+6.491389
+7.211866
+6.776532
+5.576763
+5.7659464
+6.543843
+5.57827
+6.709599
+6.8726244
+6.59173
+6.631732
+6.9985876
+6.87736
+7.489514
+5.7071085
+7.478836
+5.8673086
+6.6094766
+6.31264
+5.267159
+5.078062
+7.6211348
+6.527839
+7.064787
+6.698535
+6.037001
+5.423989
+6.3911605
+6.444887
+6.695321
+6.7394996
+5.736779
+5.216387
+7.045471
+7.211866
+5.9609942
+5.67116
+7.0122066
+6.5268397
+5.684079
+6.491335
+7.529561
+6.29732
+6.1150737
+7.3355327
+7.3476768
+6.4410257
+5.725683
+6.767765
+6.610578
+6.195107
+7.193926
+6.7086015
+6.503502
+6.126737
+6.401618
+6.2685027
+7.2798486
+7.1237426
+6.576234
+5.5889177
+5.5564
+6.631732
+6.4358587
+6.504532
+5.6332417
+6.166194
+7.407914
+6.977326
+6.3540125
+6.530521
+6.855708
+6.266541
+6.200362
+6.580424
+5.5941973
+7.3038073
+6.5268397
+7.407914
+7.648735
+6.644659
+7.3630214
+7.3830495
+6.7175026
+7.9444075
+5.551831
+6.2259016
+6.815142
+5.5941973
+7.5179734
+6.8726244
+5.418245
+7.0771327
+7.356374
+7.985815
+7.1549835
+6.7114034
+6.5329785
+7.9444075
+5.702716
+7.032158
+6.6356225
+6.4751844
+6.2157063
+7.002886
+6.3380733
+4.9186764
+7.059797
+6.512761
+5.5139256
+6.4034224
+6.8924704
+6.171549
+6.8826456
+7.283107
+6.5136123
+6.070827
+7.2054534
+6.7876177
+7.114401
+6.3176
+6.5958576
+6.9308805
+5.94277
+7.2578697
+6.7709775
+6.6093802
+7.0226016
+7.6956105
+6.57564
+5.24795
+6.5658736
+6.09532
+5.8043475
+5.1326013
+5.5831256
+5.791923
+7.2986455
+6.9308805
+6.8938518
+6.401618
+6.6153946
+7.100722
+6.6537275
+6.0852942
+6.6262474
+7.2718153
+6.8286877
+7.033388
+7.1461835
+5.4987116
+6.2271605
+6.113218
+6.5715804
+6.4013047
+5.693973
+6.3374543
+7.207704
+7.200252
+7.022265
+7.061931
+7.356374
+7.785687
+7.7303634
+6.6093802
+5.414932
+6.418995
+7.121869
+6.9924088
+6.873275
+7.2392983
+6.5481467
+6.92484
+6.8052654
+5.17515
+6.609482
+6.4053726
+7.087753
+6.554045
+5.2214603
+7.283107
+5.535821
+5.67116
+7.253158
+6.612894
+6.166194
+7.3042836
+6.184141
+7.0771327
+6.7876177
+7.287871
+7.4218545
+6.1883736
+7.0770483
+6.535317
+6.832174
+6.600105
+5.954684
+7.4687033
+6.279503
+6.3911605
+6.3022556
+5.92696
+5.5817537
+7.177708
+6.8230877
+6.130222
+6.73893
+6.597915
+6.3860326
+5.8244486
+7.4675035
+6.73782
+7.7303634
+6.576234
+5.9591675
+7.528009
+7.463671
+7.776771
+4.9186764
+7.2441607
+7.4598103
+5.555604
+5.925364
+6.055718
+6.818296
+6.9219446
+7.1237426
+6.7072544
+8.132364
+7.765772
+6.306419
+6.5526924
+6.629962
+7.407914
+6.88002
+7.4218545
+6.2322044
+5.7431617
+6.608367
+7.1270494
+5.7034698
+6.631732
+5.895033
+5.6011834
+5.2891855
+7.322918
+6.670511
+6.1788745
+5.8673086
+7.0759993
+5.709116
+6.330988
+5.2891855
+7.190937
+6.5914927
+7.9216895
+6.3681445
+7.449414
+7.2305517
+6.6684856
+6.875267
+6.856292
+6.200362
+6.194875
+7.7046857
+6.291841
+6.8455205
+7.1434317
+7.045471
+7.417041
+6.519386
+6.5465546
+5.832465
+4.9847794
+5.7858152
+6.1484227
+4.9847794
+7.002103
+7.3764777
+6.9037566
+5.7687716
+6.5935707
+5.7544155
+7.657319
+7.1032205
+6.597572
+6.5268397
+5.524064
+6.4302344
+6.9143195
+7.4307804
+6.4751844
+7.2464986
+6.4751844
+6.969932
+6.877607
+7.257072
+7.658885
+7.101384
+7.1734056
+7.180482
+7.417041
+6.795477
+6.286934
+5.5353627
+6.3492208
+7.200252
+7.1749763
+7.356374
+6.184141
+6.2948523
+7.246741
+7.418542
+7.101068
+5.7008233
+7.287871
+7.478836
+6.384309
+6.539104
+6.0423512
+6.3184676
+6.3126426
+6.515685
+7.152094
+7.6351986
+7.0226016
+6.932092
+6.0898604
+6.631732
+6.1096177
+6.200362
+6.028014
+7.0226016
+6.3851595
+5.4892125
+6.57564
+7.3830495
+6.595982
+5.2004204
+5.913258
+6.279041
+6.4515824
+7.018715
+6.737904
+5.695808
+5.3844495
+6.21219
+6.7588897
+5.564103
+7.1074014
+7.101068
+6.4452515
+6.082627
+6.360474
+5.6854997
+7.7070384
+7.3236628
+6.4016185
+6.117755
+5.3257074
+6.5914974
+5.545625
+6.1585646
+5.8972793
+5.69482
+6.581993
+6.7855964

baselines/DeepDTA/source/output/GPCR/eval_results

+RMSE : 1.4034081434737957 ; Pearson Correlation Coefficient : 0.2416971016625298
\ No newline at end of file

baselines/DeepDTA/source/output/Ion_channel/eval_results

+RMSE : 1.4512754880382956 ; Pearson Correlation Coefficient : 0.11439684637165645
\ No newline at end of file

baselines/DeepDTA/source/output/Tyrosine_kinase/eval_results

+RMSE : 1.2566115226505494 ; Pearson Correlation Coefficient : 0.3483553292295794
\ No newline at end of file

baselines/DeepDTA/source/output/test/eval_results

+RMSE : 0.9898476924424743 ; Pearson Correlation Coefficient : 0.7531926430165059
\ No newline at end of file

baselines/DeepDTA/source/run_baseline.py

+from __future__ import print_function
+#import matplotlib
+#matplotlib.use('Agg')
+import numpy as np
+import tensorflow as tf
+import random as rn
+
+### We modified Pahikkala et al. (2014) source code for cross-val process ###
+
+import os
+
+from torch import save
+os.environ["CUDA_VISIBLE_DEVICES"] = "6"
+
+os.environ['PYTHONHASHSEED'] = '0'
+
+np.random.seed(1)
+rn.seed(1)
+
+session_conf = tf.ConfigProto(intra_op_parallelism_threads=1, inter_op_parallelism_threads=1)
+import keras
+from keras import backend as K
+tf.set_random_seed(0)
+sess = tf.Session(graph=tf.get_default_graph(), config=session_conf)
+K.set_session(sess)
+
+
+from datahelper import *
+#import logging
+from itertools import product
+from arguments import argparser, logging
+
+import keras
+from keras.models import Model
+from keras.preprocessing import sequence
+from keras.models import Sequential, load_model
+from keras.layers import Dense, Dropout, Activation
+from keras.layers import Embedding
+from keras.layers import Conv1D, GlobalMaxPooling1D, MaxPooling1D
+from keras.layers.normalization import BatchNormalization
+from keras.layers import Conv2D, GRU
+from keras.layers import Input, Embedding, LSTM, Dense, TimeDistributed, Masking, RepeatVector, merge, Flatten
+from keras.models import Model
+from keras.utils import plot_model
+from keras.layers import Bidirectional
+from keras.callbacks import ModelCheckpoint, EarlyStopping
+from keras import optimizers, layers
+
+
+import sys, pickle, os
+import math, json, time
+import decimal
+import matplotlib.pyplot as plt
+import matplotlib.mlab as mlab
+from random import shuffle
+from copy import deepcopy
+from sklearn import preprocessing
+from emetrics import get_aupr, get_cindex, get_rm2
+
+
+
+TABSY = "\t"
+figdir = "figures/"
+
+def build_combined_onehot(FLAGS, NUM_FILTERS, FILTER_LENGTH1, FILTER_LENGTH2):
+    XDinput = Input(shape=(FLAGS.max_smi_len, FLAGS.charsmiset_size))
+    XTinput = Input(shape=(FLAGS.max_seq_len, FLAGS.charseqset_size))
+
+
+    encode_smiles= Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(XDinput)
+    encode_smiles = Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(encode_smiles)
+    encode_smiles = Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(encode_smiles)
+    encode_smiles = GlobalMaxPooling1D()(encode_smiles) #pool_size=pool_length[i]
+
+
+    encode_protein = Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(XTinput)
+    encode_protein = Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(encode_protein)
+    encode_protein = Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(encode_protein)
+    encode_protein = GlobalMaxPooling1D()(encode_protein)
+
+
+
+    encode_interaction = keras.layers.concatenate([encode_smiles, encode_protein])
+    #encode_interaction = keras.layers.concatenate([encode_smiles, encode_protein], axis=-1) #merge.Add()([encode_smiles, encode_protein])
+
+    # Fully connected 
+    FC1 = Dense(1024, activation='relu')(encode_interaction)
+    FC2 = Dropout(0.1)(FC1)
+    FC2 = Dense(1024, activation='relu')(FC2)
+    FC2 = Dropout(0.1)(FC2)
+    FC2 = Dense(512, activation='relu')(FC2)
+
+
+    predictions = Dense(1, kernel_initializer='normal')(FC2) 
+
+    interactionModel = Model(inputs=[XDinput, XTinput], outputs=[predictions])
+    interactionModel.compile(optimizer='adam', loss='mean_squared_error', metrics=[cindex_score]) #, metrics=['cindex_score']
+    
+
+    print(interactionModel.summary())
+    plot_model(interactionModel, to_file='figures/build_combined_onehot.png')
+
+    return interactionModel
+
+
+
+
+
+def build_combined_categorical(FLAGS, NUM_FILTERS, FILTER_LENGTH1, FILTER_LENGTH2):
+   
+    XDinput = Input(shape=(FLAGS.max_smi_len,), dtype='int32') ### Buralar flagdan gelmeliii
+    XTinput = Input(shape=(FLAGS.max_seq_len,), dtype='int32')
+
+    ### SMI_EMB_DINMS  FLAGS GELMELII 
+    encode_smiles = Embedding(input_dim=FLAGS.charsmiset_size+1, output_dim=128, input_length=FLAGS.max_smi_len)(XDinput) 
+    encode_smiles = Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(encode_smiles)
+    encode_smiles = Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(encode_smiles)
+    encode_smiles = Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(encode_smiles)
+    encode_smiles = GlobalMaxPooling1D()(encode_smiles)
+
+
+    encode_protein = Embedding(input_dim=FLAGS.charseqset_size+1, output_dim=128, input_length=FLAGS.max_seq_len)(XTinput)
+    encode_protein = Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(encode_protein)
+    encode_protein = Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(encode_protein)
+    encode_protein = Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(encode_protein)
+    encode_protein = GlobalMaxPooling1D()(encode_protein)
+
+
+    encode_interaction = keras.layers.concatenate([encode_smiles, encode_protein], axis=-1) #merge.Add()([encode_smiles, encode_protein])
+
+    # Fully connected 
+    FC1 = Dense(1024, activation='relu')(encode_interaction)
+    FC2 = Dropout(0.1)(FC1)
+    FC2 = Dense(1024, activation='relu')(FC2)
+    FC2 = Dropout(0.1)(FC2)
+    FC2 = Dense(512, activation='relu')(FC2)
+
+
+    # And add a logistic regression on top
+    predictions = Dense(1, kernel_initializer='normal')(FC2) #OR no activation, rght now it's between 0-1, do I want this??? activation='sigmoid'
+
+    interactionModel = Model(inputs=[XDinput, XTinput], outputs=[predictions])
+
+    interactionModel.compile(optimizer='adam', loss='mean_squared_error', metrics=[cindex_score]) #, metrics=['cindex_score']
+    print(interactionModel.summary())
+    plot_model(interactionModel, to_file='figures/build_combined_categorical.png')
+
+    return interactionModel
+
+
+
+def build_single_drug(FLAGS, NUM_FILTERS, FILTER_LENGTH1, FILTER_LENGTH2):
+   
+    interactionModel = Sequential()
+    XTmodel = Sequential()
+    XTmodel.add(Activation('linear', input_shape=(FLAGS.target_count,)))
+
+
+    encode_smiles = Sequential()
+    encode_smiles.add(Embedding(input_dim=FLAGS.charsmiset_size+1, output_dim=128, input_length=FLAGS.max_smi_len)) 
+    encode_smiles.add(Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)) #input_shape=(MAX_SMI_LEN, SMI_EMBEDDING_DIMS)
+    encode_smiles.add(Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1))
+    encode_smiles.add(Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1))
+    encode_smiles.add(GlobalMaxPooling1D())
+
+
+    interactionModel.add(Merge([encode_smiles, XTmodel], mode='concat', concat_axis=1))
+    #interactionModel.add(layers.merge.Concatenate([XDmodel, XTmodel]))
+
+    # Fully connected 
+    interactionModel.add(Dense(1024, activation='relu')) #1024
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(1024, activation='relu')) #1024
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(512, activation='relu')) 
+
+
+    interactionModel.add(Dense(1, kernel_initializer='normal'))
+    interactionModel.compile(optimizer='adam', loss='mean_squared_error', metrics=[cindex_score])
+
+    print(interactionModel.summary())
+    plot_model(interactionModel, to_file='figures/build_single_drug.png')
+
+    return interactionModel
+
+
+def build_single_prot(FLAGS, NUM_FILTERS, FILTER_LENGTH1, FILTER_LENGTH2):
+   
+    interactionModel = Sequential()
+    XDmodel = Sequential()
+    XDmodel.add(Activation('linear', input_shape=(FLAGS.drugcount,)))
+
+
+    XTmodel1 = Sequential()
+    XTmodel1.add(Embedding(input_dim=FLAGS.charseqset_size+1, output_dim=128,  input_length=FLAGS.max_seq_len))
+    XTmodel1.add(Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)) #input_shape=(MAX_SEQ_LEN, SEQ_EMBEDDING_DIMS)
+    XTmodel1.add(Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1))
+    XTmodel1.add(Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1))
+    XTmodel1.add(GlobalMaxPooling1D())
+
+
+    interactionModel.add(Merge([XDmodel, XTmodel1], mode='concat', concat_axis=1))
+
+    # Fully connected 
+    interactionModel.add(Dense(1024, activation='relu'))
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(1024, activation='relu'))
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(512, activation='relu'))
+
+    interactionModel.add(Dense(1, kernel_initializer='normal'))
+    interactionModel.compile(optimizer='adam', loss='mean_squared_error', metrics=[cindex_score])
+
+    print(interactionModel.summary())
+    plot_model(interactionModel, to_file='figures/build_single_protein.png')
+
+    return interactionModel
+
+def build_baseline(FLAGS, NUM_FILTERS, FILTER_LENGTH1, FILTER_LENGTH2):
+    interactionModel = Sequential()
+
+    XDmodel = Sequential()
+    XDmodel.add(Dense(1, activation='linear', input_shape=(FLAGS.drug_count, )))
+
+    XTmodel = Sequential()
+    XTmodel.add(Dense(1, activation='linear', input_shape=(FLAGS.target_count,)))
+
+
+    interactionModel.add(Merge([XDmodel, XTmodel], mode='concat', concat_axis=1))
+
+    # Fully connected 
+    interactionModel.add(Dense(1024, activation='relu'))
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(1024, activation='relu'))
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(512, activation='relu'))
+
+    interactionModel.add(Dense(1, kernel_initializer='normal'))
+    interactionModel.compile(optimizer='adam', loss='mean_squared_error', metrics=[cindex_score])
+
+    print(interactionModel.summary())
+    plot_model(interactionModel, to_file='figures/build_baseline.png')
+
+    return interactionModel
+
+def nfold_1_2_3_setting_sample(XD, XT,  Y, measure, runmethod,  FLAGS, dataset):
+
+    bestparamlist = []
+    # test_set, outer_train_sets = dataset.read_sets(FLAGS) 
+    
+    # foldinds = len(outer_train_sets)
+
+    # test_sets = []
+    # ## TRAIN AND VAL
+    # val_sets = []
+    # train_sets = []
+
+    #logger.info('Start training')
+    # for val_foldind in range(foldinds):
+    #     val_fold = outer_train_sets[val_foldind]
+    #     val_sets.append(val_fold)
+    #     otherfolds = deepcopy(outer_train_sets)
+    #     otherfolds.pop(val_foldind)
+    #     otherfoldsinds = [item for sublist in otherfolds for item in sublist]
+    #     train_sets.append(otherfoldsinds)
+    #     test_sets.append(test_set)
+    #     print("val set", str(len(val_fold)))
+    #     print("train set", str(len(otherfoldsinds)))
+
+
+
+    # bestparamind, best_param_list, bestperf, all_predictions_not_need, losses_not_need = general_nfold_cv(XD, XT,  Y, label_row_inds, label_col_inds, 
+    #                                                                                             measure, runmethod, FLAGS, train_sets, val_sets)
+   
+    #print("Test Set len", str(len(test_set)))
+    #print("Outer Train Set len", str(len(outer_train_sets)))
+    # bestparam, best_param_list, bestperf, all_predictions, all_losses = general_nfold_cv(XD, XT,  Y, label_row_inds, label_col_inds, 
+    #                                                                                             measure, runmethod, FLAGS, train_sets, test_sets)
+    # bestparam, best_param_list, bestperf, all_predictions, all_losses = 
+    general_nfold_cv(XD, XT,  Y, measure, runmethod, FLAGS)
+    # testperf = all_predictions[bestparamind]##pointer pos 
+
+    # logging("---FINAL RESULTS-----", FLAGS)
+    # logging("best param index = %s,  best param = %.5f" % 
+    #         (bestparamind, bestparam), FLAGS)
+
+
+    # testperfs = []
+    # testloss= []
+
+    # avgperf = 0.
+
+    # for test_foldind in range(len(test_sets)):
+    #     foldperf = all_predictions[bestparamind][test_foldind]
+    #     foldloss = all_losses[bestparamind][test_foldind]
+    #     testperfs.append(foldperf)
+    #     testloss.append(foldloss)
+    #     avgperf += foldperf
+
+    # avgperf = avgperf / len(test_sets)
+    # avgloss = np.mean(testloss)
+    # teststd = np.std(testperfs)
+
+    # logging("Test Performance CI", FLAGS)
+    # logging(testperfs, FLAGS)
+    # logging("Test Performance MSE", FLAGS)
+    # logging(testloss, FLAGS)
+
+    # return avgperf, avgloss, teststd
+
+
+
+
+def general_nfold_cv(XD, XT,  Y, prfmeasure, runmethod, FLAGS): ## BURAYA DA FLAGS LAZIM????
+    
+    paramset1 = FLAGS.num_windows                              #[32]#[32,  512] #[32, 128]  # filter numbers
+    paramset2 = FLAGS.smi_window_lengths                               #[4, 8]#[4,  32] #[4,  8] #filter length smi
+    paramset3 = FLAGS.seq_window_lengths                               #[8, 12]#[64,  256] #[64, 192]#[8, 192, 384]
+    epoch = FLAGS.num_epoch                                 #100
+    batchsz = FLAGS.batch_size                             #256
+
+    logging("---Parameter Search-----", FLAGS)
+
+    # w = len(val_sets)
+    # h = len(paramset1) * len(paramset2) * len(paramset3)
+
+    # all_predictions = [[0 for x in range(w)] for y in range(h)] 
+    # all_losses = [[0 for x in range(w)] for y in range(h)] 
+    # print(all_predictions)
+
+    # for foldind in range(len(val_sets)):
+    #     valinds = val_sets[foldind]
+    #     labeledinds = labeled_sets[foldind]
+
+    #     Y_train = np.mat(np.copy(Y))
+
+    #     params = {}
+    #     XD_train = XD
+    #     XT_train = XT
+    #     trrows = label_row_inds[labeledinds]
+    #     trcols = label_col_inds[labeledinds]
+
+    #     XD_train = XD[trrows]
+    #     XT_train = XT[trcols]
+
+    #     train_drugs, train_prots,  train_Y = prepare_interaction_pairs(XD, XT, Y, trrows, trcols)
+        
+    #     terows = label_row_inds[valinds]
+    #     tecols = label_col_inds[valinds]
+    #     #print("terows", str(terows), str(len(terows)))
+    #     #print("tecols", str(tecols), str(len(tecols)))
+
+    #     val_drugs, val_prots,  val_Y = prepare_interaction_pairs(XD, XT,  Y, terows, tecols)
+
+
+    pointer = 0
+       
+    for param1ind in range(len(paramset1)): #hidden neurons
+        param1value = paramset1[param1ind]
+        for param2ind in range(len(paramset2)): #learning rate
+            param2value = paramset2[param2ind]
+
+            for param3ind in range(len(paramset3)):
+                param3value = paramset3[param3ind]
+
+                gridmodel = runmethod(FLAGS, param1value, param2value, param3value)
+                es = EarlyStopping(monitor='loss', mode='min', verbose=1, patience=15)
+                gridres = gridmodel.fit(([np.array(XD),np.array(XT)]), np.array(Y), batch_size=batchsz, epochs=epoch, 
+                        # validation_data=( ([np.array(val_drugs), np.array(val_prots) ]), np.array(val_Y)), 
+                        shuffle=False, 
+                        callbacks=[es]) 
+                
+                gridmodel.save('model/my_model_{}'.format(pointer))
+
+                # predicted_labels = gridmodel.predict([np.array(val_drugs), np.array(val_prots) ])
+                # loss, rperf2 = gridmodel.evaluate(([np.array(val_drugs),np.array(val_prots) ]), np.array(val_Y), verbose=0)
+                # rperf = prfmeasure(val_Y, predicted_labels)
+                # rperf = rperf[0]
+
+
+                # logging("P1 = %d,  P2 = %d, P3 = %d, Fold = %d, CI-i = %f, CI-ii = %f, MSE = %f" % 
+                # (param1ind, param2ind, param3ind, foldind, rperf, rperf2, loss), FLAGS)
+
+                plotLoss(gridres, param1ind, param2ind, param3ind)
+
+                # all_predictions[pointer][foldind] =rperf #TODO FOR EACH VAL SET allpredictions[pointer][foldind]
+                # all_losses[pointer][foldind]= loss
+
+                pointer +=1
+
+    # bestperf = -float('Inf')
+    # bestpointer = None
+
+
+    # best_param_list = []
+    # ##Take average according to folds, then chooose best params
+    # pointer = 0
+    # for param1ind in range(len(paramset1)):
+    #         for param2ind in range(len(paramset2)):
+    #             for param3ind in range(len(paramset3)):
+                
+    #                 avgperf = 0.
+    #                 for foldind in range(len(val_sets)):
+    #                     foldperf = all_predictions[pointer][foldind]
+    #                     avgperf += foldperf
+    #                 avgperf /= len(val_sets)
+    #                 #print(epoch, batchsz, avgperf)
+    #                 if avgperf > bestperf:
+    #                     bestperf = avgperf
+    #                     bestpointer = pointer
+    #                     best_param_list = [param1ind, param2ind, param3ind]
+
+    #                 pointer +=1
+        
+    # return  bestpointer, best_param_list, bestperf, all_predictions, all_losses
+
+
+
+def cindex_score(y_true, y_pred):
+
+    g = tf.subtract(tf.expand_dims(y_pred, -1), y_pred)
+    g = tf.cast(g == 0.0, tf.float32) * 0.5 + tf.cast(g > 0.0, tf.float32)
+
+    f = tf.subtract(tf.expand_dims(y_true, -1), y_true) > 0.0
+    f = tf.matrix_band_part(tf.cast(f, tf.float32), -1, 0)
+
+    g = tf.reduce_sum(tf.multiply(g, f))
+    f = tf.reduce_sum(f)
+
+    return tf.where(tf.equal(g, 0), 0.0, g/f) #select
+
+
+   
+def plotLoss(history, batchind, epochind, param3ind):
+
+    figname = "b"+str(batchind) + "_e" + str(epochind) + "_" + str(param3ind) + "_"  + "_" + str(time.time()) 
+    plt.figure()
+    plt.plot(history.history['loss'])
+    # plt.plot(history.history['val_loss'])
+    plt.title('model loss')
+    plt.ylabel('loss')
+    plt.xlabel('epoch')
+	#plt.legend(['trainloss', 'valloss', 'cindex', 'valcindex'], loc='upper left')
+    plt.legend(['trainloss'], loc='upper left')
+    plt.savefig("figures/"+figname +".png" , dpi=None, facecolor='w', edgecolor='w', orientation='portrait', 
+                    papertype=None, format=None,transparent=False, bbox_inches=None, pad_inches=0.1,frameon=None)
+    plt.close()
+
+
+    ## PLOT CINDEX
+    plt.figure()
+    plt.title('model concordance index') 
+    plt.ylabel('cindex')
+    plt.xlabel('epoch')
+    plt.plot(history.history['cindex_score'])
+    # plt.plot(history.history['val_cindex_score'])
+    plt.legend(['traincindex'], loc='upper left')
+    plt.savefig("figures/"+figname + "_acc.png" , dpi=None, facecolor='w', edgecolor='w', orientation='portrait', 
+                            papertype=None, format=None,transparent=False, bbox_inches=None, pad_inches=0.1,frameon=None)
+    plt.close()
+
+
+
+def prepare_interaction_pairs(XD, XT,  Y, rows, cols):
+    drugs = []
+    targets = []
+    targetscls = []
+    affinity=[] 
+        
+    for pair_ind in range(len(rows)):
+        drug = XD[rows[pair_ind]]
+        drugs.append(drug)
+
+        target=XT[cols[pair_ind]]
+        targets.append(target)
+
+        affinity.append(Y[rows[pair_ind],cols[pair_ind]])
+
+    drug_data = np.stack(drugs)
+    target_data = np.stack(targets)
+
+    return drug_data,target_data,  affinity
+
+
+       
+def experiment(FLAGS, perfmeasure, deepmethod, foldcount=6): #5-fold cross validation + test
+
+    #Input
+    #XD: [drugs, features] sized array (features may also be similarities with other drugs
+    #XT: [targets, features] sized array (features may also be similarities with other targets
+    #Y: interaction values, can be real values or binary (+1, -1), insert value float("nan") for unknown entries
+    #perfmeasure: function that takes as input a list of correct and predicted outputs, and returns performance
+    #higher values should be better, so if using error measures use instead e.g. the inverse -error(Y, P)
+    #foldcount: number of cross-validation folds for settings 1-3, setting 4 always runs 3x3 cross-validation
+
+
+    dataset = DataSet_for_new( fpath = FLAGS.dataset_path, ### BUNU ARGS DA GUNCELLE
+                      setting_no = FLAGS.problem_type, ##BUNU ARGS A EKLE
+                      seqlen = FLAGS.max_seq_len,
+                      smilen = FLAGS.max_smi_len,
+                      need_shuffle = False )
+    # set character set size
+    FLAGS.charseqset_size = dataset.charseqset_size 
+    FLAGS.charsmiset_size = dataset.charsmiset_size 
+
+    XD, XT, Y = dataset.parse_data(FLAGS)
+
+    XD = np.asarray(XD)
+    XT = np.asarray(XT)
+    Y = np.asarray(Y)
+
+    drugcount = XD.shape[0]
+    print(drugcount)
+    targetcount = XT.shape[0]
+    print(targetcount)
+
+    FLAGS.drug_count = drugcount
+    FLAGS.target_count = targetcount
+
+    # label_row_inds, label_col_inds = np.where(np.isnan(Y)==False)  #basically finds the point address of affinity [x,y]
+
+    if not os.path.exists(figdir):
+        os.makedirs(figdir)
+
+    print(FLAGS.log_dir)
+    # S1_avgperf, S1_avgloss, S1_teststd = 
+    nfold_1_2_3_setting_sample(XD, XT, Y, perfmeasure, deepmethod, FLAGS, dataset)
+
+    logging("Setting " + str(FLAGS.problem_type), FLAGS)
+    # logging("avg_perf = %.5f,  avg_mse = %.5f, std = %.5f" % 
+    #         (S1_avgperf, S1_avgloss, S1_teststd), FLAGS)
+
+def test_experiment(FLAGS, perfmeasure, deepmethod, foldcount=6): #5-fold cross validation + test
+
+    #Input
+    #XD: [drugs, features] sized array (features may also be similarities with other drugs
+    #XT: [targets, features] sized array (features may also be similarities with other targets
+    #Y: interaction values, can be real values or binary (+1, -1), insert value float("nan") for unknown entries
+    #perfmeasure: function that takes as input a list of correct and predicted outputs, and returns performance
+    #higher values should be better, so if using error measures use instead e.g. the inverse -error(Y, P)
+    #foldcount: number of cross-validation folds for settings 1-3, setting 4 always runs 3x3 cross-validation
+
+
+    dataset = DataSet_for_new( fpath = FLAGS.dataset_path, ### BUNU ARGS DA GUNCELLE
+                      setting_no = FLAGS.problem_type, ##BUNU ARGS A EKLE
+                      seqlen = FLAGS.max_seq_len,
+                      smilen = FLAGS.max_smi_len,
+                      need_shuffle = False )
+    # set character set size
+    FLAGS.charseqset_size = dataset.charseqset_size 
+    FLAGS.charsmiset_size = dataset.charsmiset_size 
+
+    XD, XT, Y = dataset.parse_data(FLAGS)
+
+    XD = np.asarray(XD)
+    XT = np.asarray(XT)
+    Y = np.asarray(Y)
+
+    drugcount = XD.shape[0]
+    print(drugcount)
+    targetcount = XT.shape[0]
+    print(targetcount)
+
+    FLAGS.drug_count = drugcount
+    FLAGS.target_count = targetcount
+
+    # label_row_inds, label_col_inds = np.where(np.isnan(Y)==False)  #basically finds the point address of affinity [x,y]
+
+    if not os.path.exists(FLAGS.out):
+        os.makedirs(FLAGS.out)
+
+    print(FLAGS.out)
+    # drug_data = np.stack(drugs)
+    # target_data = np.stack(targets)
+    
+    gridmodel = keras.models.load_model(FLAGS.model, compile=False)
+    predicted_labels = gridmodel.predict([np.array(XD), np.array(XT) ])
+    save_path = FLAGS.out + "results.txt"
+    with open(save_path, "w") as f:
+        for i in predicted_labels:
+            f.write(str(i[0]) + '\n')
+    print("---------------------predict over-------------------------")
+    # S1_avgperf, S1_avgloss, S1_teststd = 
+    # nfold_1_2_3_setting_sample(XD, XT, Y, perfmeasure, deepmethod, FLAGS, dataset)
+
+    # logging("Setting " + str(FLAGS.problem_type), FLAGS)
+    # logging("avg_perf = %.5f,  avg_mse = %.5f, std = %.5f" % 
+    #         (S1_avgperf, S1_avgloss, S1_teststd), FLAGS)
+
+
+
+
+
+def run_regression( FLAGS ): 
+
+    perfmeasure = get_cindex
+    deepmethod = build_combined_categorical
+
+    experiment(FLAGS, perfmeasure, deepmethod)
+
+
+
+def run_predict(FLAGS):
+    perfmeasure = get_cindex
+    deepmethod = build_combined_categorical
+    test_experiment(FLAGS, perfmeasure, deepmethod)
+
+
+if __name__=="__main__":
+    
+    FLAGS = argparser()
+    
+    
+    # train
+    # FLAGS.num_windows = [32]
+    # FLAGS.seq_window_lengths = [8, 12]
+    # FLAGS.smi_window_lengths = [4, 8]
+    # FLAGS.batch_size = 256
+    # FLAGS.num_epoch = 2
+    # FLAGS.max_seq_len = 1000
+    # FLAGS.max_smi_len = 100
+    # # FLAGS.dataset_path = 'data/kiba/'
+    # FLAGS.dataset_path = {"ligand":'../../../data/train/train_smile', 
+    #                       "protein":'../../../data/train/train_protein_seq',
+    #                       "y":'../../../data/train/train_ic50',}
+    # FLAGS.problem_type = 1
+    # FLAGS.log_dir = "logs/"
+    
+    # FLAGS.log_dir = FLAGS.log_dir + str(time.time()) + "/"
+
+    
+    # if not os.path.exists(FLAGS.log_dir):
+    #     os.makedirs(FLAGS.log_dir)
+
+    # logging(str(FLAGS), FLAGS)
+    # run_regression( FLAGS )
+
+    
+    # test
+    FLAGS.max_seq_len = 1000
+    FLAGS.max_smi_len = 100
+    # FLAGS.dataset_path = {"ligand":'../../../data/test/test_smile', 
+    #                       "protein":'../../../data/test/test_protein_seq',
+    #                       "y":'../../../data/test/test_ic50'}
+    # FLAGS.dataset_path = {"ligand":'../../../data/ER/ER_smile', 
+    #                       "protein":'../../../data/ER/ER_protein_seq',
+    #                       "y":'../../../data/ER/ER_ic50'}
+    # FLAGS.dataset_path = {"ligand":'../../../data/GPCR/GPCR_smile', 
+    #                       "protein":'../../../data/GPCR/GPCR_protein_seq',
+    #                       "y":'../../../data/GPCR/GPCR_ic50'}
+    # FLAGS.dataset_path = {"ligand":'../../../data/Ion_channel/channel_smile',
+    #                       "protein":'../../../data/Ion_channel/channel_protein_seq',
+    #                       "y":'../../../data/Ion_channel/channel_ic50'}
+    FLAGS.dataset_path = {"ligand":'../../../data/Tyrosine_kinase/kinase_smile',
+                          "protein":'../../../data/Tyrosine_kinase/kinase_protein_seq',
+                          "y":'../../../data/Tyrosine_kinase/kinase_ic50'}
+    FLAGS.out = "./output/Tyrosine_kinase/"
+    FLAGS.model = "model/my_model_0.h5"
+    run_predict(FLAGS)
\ No newline at end of file

baselines/DeepDTA/source/run_experiments.py

+from __future__ import print_function
+#import matplotlib
+#matplotlib.use('Agg')
+import numpy as np
+import tensorflow as tf
+import random as rn
+
+### We modified Pahikkala et al. (2014) source code for cross-val process ###
+
+import os
+os.environ["CUDA_VISIBLE_DEVICES"] = "3"
+
+os.environ['PYTHONHASHSEED'] = '0'
+
+np.random.seed(1)
+rn.seed(1)
+
+session_conf = tf.ConfigProto(intra_op_parallelism_threads=1, inter_op_parallelism_threads=1)
+import keras
+from keras import backend as K
+tf.set_random_seed(0)
+sess = tf.Session(graph=tf.get_default_graph(), config=session_conf)
+K.set_session(sess)
+
+
+from datahelper import *
+#import logging
+from itertools import product
+from arguments import argparser, logging
+
+import keras
+from keras.models import Model
+from keras.preprocessing import sequence
+from keras.models import Sequential, load_model
+from keras.layers import Dense, Dropout, Activation
+from keras.layers import Embedding
+from keras.layers import Conv1D, GlobalMaxPooling1D, MaxPooling1D
+from keras.layers.normalization import BatchNormalization
+from keras.layers import Conv2D, GRU
+from keras.layers import Input, Embedding, LSTM, Dense, TimeDistributed, Masking, RepeatVector, merge, Flatten
+from keras.models import Model
+from keras.utils import plot_model
+from keras.layers import Bidirectional
+from keras.callbacks import ModelCheckpoint, EarlyStopping
+from keras import optimizers, layers
+
+
+import sys, pickle, os
+import math, json, time
+import decimal
+import matplotlib.pyplot as plt
+import matplotlib.mlab as mlab
+from random import shuffle
+from copy import deepcopy
+from sklearn import preprocessing
+from emetrics import get_aupr, get_cindex, get_rm2
+
+
+
+TABSY = "\t"
+figdir = "figures/"
+
+def build_combined_onehot(FLAGS, NUM_FILTERS, FILTER_LENGTH1, FILTER_LENGTH2):
+    XDinput = Input(shape=(FLAGS.max_smi_len, FLAGS.charsmiset_size))
+    XTinput = Input(shape=(FLAGS.max_seq_len, FLAGS.charseqset_size))
+
+
+    encode_smiles= Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(XDinput)
+    encode_smiles = Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(encode_smiles)
+    encode_smiles = Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(encode_smiles)
+    encode_smiles = GlobalMaxPooling1D()(encode_smiles) #pool_size=pool_length[i]
+
+
+    encode_protein = Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(XTinput)
+    encode_protein = Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(encode_protein)
+    encode_protein = Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(encode_protein)
+    encode_protein = GlobalMaxPooling1D()(encode_protein)
+
+
+
+    encode_interaction = keras.layers.concatenate([encode_smiles, encode_protein])
+    #encode_interaction = keras.layers.concatenate([encode_smiles, encode_protein], axis=-1) #merge.Add()([encode_smiles, encode_protein])
+
+    # Fully connected 
+    FC1 = Dense(1024, activation='relu')(encode_interaction)
+    FC2 = Dropout(0.1)(FC1)
+    FC2 = Dense(1024, activation='relu')(FC2)
+    FC2 = Dropout(0.1)(FC2)
+    FC2 = Dense(512, activation='relu')(FC2)
+
+
+    predictions = Dense(1, kernel_initializer='normal')(FC2) 
+
+    interactionModel = Model(inputs=[XDinput, XTinput], outputs=[predictions])
+    interactionModel.compile(optimizer='adam', loss='mean_squared_error', metrics=[cindex_score]) #, metrics=['cindex_score']
+    
+
+    print(interactionModel.summary())
+    plot_model(interactionModel, to_file='figures/build_combined_onehot.png')
+
+    return interactionModel
+
+
+
+
+
+def build_combined_categorical(FLAGS, NUM_FILTERS, FILTER_LENGTH1, FILTER_LENGTH2):
+   
+    XDinput = Input(shape=(FLAGS.max_smi_len,), dtype='int32') ### Buralar flagdan gelmeliii
+    XTinput = Input(shape=(FLAGS.max_seq_len,), dtype='int32')
+
+    ### SMI_EMB_DINMS  FLAGS GELMELII 
+    encode_smiles = Embedding(input_dim=FLAGS.charsmiset_size+1, output_dim=128, input_length=FLAGS.max_smi_len)(XDinput) 
+    encode_smiles = Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(encode_smiles)
+    encode_smiles = Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(encode_smiles)
+    encode_smiles = Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)(encode_smiles)
+    encode_smiles = GlobalMaxPooling1D()(encode_smiles)
+
+
+    encode_protein = Embedding(input_dim=FLAGS.charseqset_size+1, output_dim=128, input_length=FLAGS.max_seq_len)(XTinput)
+    encode_protein = Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(encode_protein)
+    encode_protein = Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(encode_protein)
+    encode_protein = Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)(encode_protein)
+    encode_protein = GlobalMaxPooling1D()(encode_protein)
+
+
+    encode_interaction = keras.layers.concatenate([encode_smiles, encode_protein], axis=-1) #merge.Add()([encode_smiles, encode_protein])
+
+    # Fully connected 
+    FC1 = Dense(1024, activation='relu')(encode_interaction)
+    FC2 = Dropout(0.1)(FC1)
+    FC2 = Dense(1024, activation='relu')(FC2)
+    FC2 = Dropout(0.1)(FC2)
+    FC2 = Dense(512, activation='relu')(FC2)
+
+
+    # And add a logistic regression on top
+    predictions = Dense(1, kernel_initializer='normal')(FC2) #OR no activation, rght now it's between 0-1, do I want this??? activation='sigmoid'
+
+    interactionModel = Model(inputs=[XDinput, XTinput], outputs=[predictions])
+
+    interactionModel.compile(optimizer='adam', loss='mean_squared_error', metrics=[cindex_score]) #, metrics=['cindex_score']
+    print(interactionModel.summary())
+    plot_model(interactionModel, to_file='figures/build_combined_categorical.png')
+
+    return interactionModel
+
+
+
+def build_single_drug(FLAGS, NUM_FILTERS, FILTER_LENGTH1, FILTER_LENGTH2):
+   
+    interactionModel = Sequential()
+    XTmodel = Sequential()
+    XTmodel.add(Activation('linear', input_shape=(FLAGS.target_count,)))
+
+
+    encode_smiles = Sequential()
+    encode_smiles.add(Embedding(input_dim=FLAGS.charsmiset_size+1, output_dim=128, input_length=FLAGS.max_smi_len)) 
+    encode_smiles.add(Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1)) #input_shape=(MAX_SMI_LEN, SMI_EMBEDDING_DIMS)
+    encode_smiles.add(Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1))
+    encode_smiles.add(Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH1,  activation='relu', padding='valid',  strides=1))
+    encode_smiles.add(GlobalMaxPooling1D())
+
+
+    interactionModel.add(Merge([encode_smiles, XTmodel], mode='concat', concat_axis=1))
+    #interactionModel.add(layers.merge.Concatenate([XDmodel, XTmodel]))
+
+    # Fully connected 
+    interactionModel.add(Dense(1024, activation='relu')) #1024
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(1024, activation='relu')) #1024
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(512, activation='relu')) 
+
+
+    interactionModel.add(Dense(1, kernel_initializer='normal'))
+    interactionModel.compile(optimizer='adam', loss='mean_squared_error', metrics=[cindex_score])
+
+    print(interactionModel.summary())
+    plot_model(interactionModel, to_file='figures/build_single_drug.png')
+
+    return interactionModel
+
+
+def build_single_prot(FLAGS, NUM_FILTERS, FILTER_LENGTH1, FILTER_LENGTH2):
+   
+    interactionModel = Sequential()
+    XDmodel = Sequential()
+    XDmodel.add(Activation('linear', input_shape=(FLAGS.drugcount,)))
+
+
+    XTmodel1 = Sequential()
+    XTmodel1.add(Embedding(input_dim=FLAGS.charseqset_size+1, output_dim=128,  input_length=FLAGS.max_seq_len))
+    XTmodel1.add(Conv1D(filters=NUM_FILTERS, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1)) #input_shape=(MAX_SEQ_LEN, SEQ_EMBEDDING_DIMS)
+    XTmodel1.add(Conv1D(filters=NUM_FILTERS*2, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1))
+    XTmodel1.add(Conv1D(filters=NUM_FILTERS*3, kernel_size=FILTER_LENGTH2,  activation='relu', padding='valid',  strides=1))
+    XTmodel1.add(GlobalMaxPooling1D())
+
+
+    interactionModel.add(Merge([XDmodel, XTmodel1], mode='concat', concat_axis=1))
+
+    # Fully connected 
+    interactionModel.add(Dense(1024, activation='relu'))
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(1024, activation='relu'))
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(512, activation='relu'))
+
+    interactionModel.add(Dense(1, kernel_initializer='normal'))
+    interactionModel.compile(optimizer='adam', loss='mean_squared_error', metrics=[cindex_score])
+
+    print(interactionModel.summary())
+    plot_model(interactionModel, to_file='figures/build_single_protein.png')
+
+    return interactionModel
+
+def build_baseline(FLAGS, NUM_FILTERS, FILTER_LENGTH1, FILTER_LENGTH2):
+    interactionModel = Sequential()
+
+    XDmodel = Sequential()
+    XDmodel.add(Dense(1, activation='linear', input_shape=(FLAGS.drug_count, )))
+
+    XTmodel = Sequential()
+    XTmodel.add(Dense(1, activation='linear', input_shape=(FLAGS.target_count,)))
+
+
+    interactionModel.add(Merge([XDmodel, XTmodel], mode='concat', concat_axis=1))
+
+    # Fully connected 
+    interactionModel.add(Dense(1024, activation='relu'))
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(1024, activation='relu'))
+    interactionModel.add(Dropout(0.1))
+    interactionModel.add(Dense(512, activation='relu'))
+
+    interactionModel.add(Dense(1, kernel_initializer='normal'))
+    interactionModel.compile(optimizer='adam', loss='mean_squared_error', metrics=[cindex_score])
+
+    print(interactionModel.summary())
+    plot_model(interactionModel, to_file='figures/build_baseline.png')
+
+    return interactionModel
+
+def nfold_1_2_3_setting_sample(XD, XT,  Y, label_row_inds, label_col_inds, measure, runmethod,  FLAGS, dataset):
+
+    bestparamlist = []
+    test_set, outer_train_sets = dataset.read_sets(FLAGS) 
+    
+    foldinds = len(outer_train_sets)
+
+    test_sets = []
+    ## TRAIN AND VAL
+    val_sets = []
+    train_sets = []
+
+    #logger.info('Start training')
+    for val_foldind in range(foldinds):
+        val_fold = outer_train_sets[val_foldind]
+        val_sets.append(val_fold)
+        otherfolds = deepcopy(outer_train_sets)
+        otherfolds.pop(val_foldind)
+        otherfoldsinds = [item for sublist in otherfolds for item in sublist]
+        train_sets.append(otherfoldsinds)
+        test_sets.append(test_set)
+        print("val set", str(len(val_fold)))
+        print("train set", str(len(otherfoldsinds)))
+
+
+
+    bestparamind, best_param_list, bestperf, all_predictions_not_need, losses_not_need = general_nfold_cv(XD, XT,  Y, label_row_inds, label_col_inds, 
+                                                                                                measure, runmethod, FLAGS, train_sets, val_sets)
+   
+    #print("Test Set len", str(len(test_set)))
+    #print("Outer Train Set len", str(len(outer_train_sets)))
+    bestparam, best_param_list, bestperf, all_predictions, all_losses = general_nfold_cv(XD, XT,  Y, label_row_inds, label_col_inds, 
+                                                                                                measure, runmethod, FLAGS, train_sets, test_sets)
+    
+    testperf = all_predictions[bestparamind]##pointer pos 
+
+    logging("---FINAL RESULTS-----", FLAGS)
+    logging("best param index = %s,  best param = %.5f" % 
+            (bestparamind, bestparam), FLAGS)
+
+
+    testperfs = []
+    testloss= []
+
+    avgperf = 0.
+
+    for test_foldind in range(len(test_sets)):
+        foldperf = all_predictions[bestparamind][test_foldind]
+        foldloss = all_losses[bestparamind][test_foldind]
+        testperfs.append(foldperf)
+        testloss.append(foldloss)
+        avgperf += foldperf
+
+    avgperf = avgperf / len(test_sets)
+    avgloss = np.mean(testloss)
+    teststd = np.std(testperfs)
+
+    logging("Test Performance CI", FLAGS)
+    logging(testperfs, FLAGS)
+    logging("Test Performance MSE", FLAGS)
+    logging(testloss, FLAGS)
+
+    return avgperf, avgloss, teststd
+
+
+
+
+def general_nfold_cv(XD, XT,  Y, label_row_inds, label_col_inds, prfmeasure, runmethod, FLAGS, labeled_sets, val_sets): ## BURAYA DA FLAGS LAZIM????
+    
+    paramset1 = FLAGS.num_windows                              #[32]#[32,  512] #[32, 128]  # filter numbers
+    paramset2 = FLAGS.smi_window_lengths                               #[4, 8]#[4,  32] #[4,  8] #filter length smi
+    paramset3 = FLAGS.seq_window_lengths                               #[8, 12]#[64,  256] #[64, 192]#[8, 192, 384]
+    epoch = FLAGS.num_epoch                                 #100
+    batchsz = FLAGS.batch_size                             #256
+
+    logging("---Parameter Search-----", FLAGS)
+
+    w = len(val_sets)
+    h = len(paramset1) * len(paramset2) * len(paramset3)
+
+    all_predictions = [[0 for x in range(w)] for y in range(h)] 
+    all_losses = [[0 for x in range(w)] for y in range(h)] 
+    print(all_predictions)
+
+    for foldind in range(len(val_sets)):
+        valinds = val_sets[foldind]
+        labeledinds = labeled_sets[foldind]
+
+        Y_train = np.mat(np.copy(Y))
+
+        params = {}
+        XD_train = XD
+        XT_train = XT
+        trrows = label_row_inds[labeledinds]
+        trcols = label_col_inds[labeledinds]
+
+        XD_train = XD[trrows]
+        XT_train = XT[trcols]
+
+        train_drugs, train_prots,  train_Y = prepare_interaction_pairs(XD, XT, Y, trrows, trcols)
+        
+        terows = label_row_inds[valinds]
+        tecols = label_col_inds[valinds]
+        #print("terows", str(terows), str(len(terows)))
+        #print("tecols", str(tecols), str(len(tecols)))
+
+        val_drugs, val_prots,  val_Y = prepare_interaction_pairs(XD, XT,  Y, terows, tecols)
+
+
+        pointer = 0
+       
+        for param1ind in range(len(paramset1)): #hidden neurons
+            param1value = paramset1[param1ind]
+            for param2ind in range(len(paramset2)): #learning rate
+                param2value = paramset2[param2ind]
+
+                for param3ind in range(len(paramset3)):
+                    param3value = paramset3[param3ind]
+
+                    gridmodel = runmethod(FLAGS, param1value, param2value, param3value)
+                    es = EarlyStopping(monitor='val_loss', mode='min', verbose=1, patience=15)
+                    gridres = gridmodel.fit(([np.array(train_drugs),np.array(train_prots) ]), np.array(train_Y), batch_size=batchsz, epochs=epoch, 
+                            validation_data=( ([np.array(val_drugs), np.array(val_prots) ]), np.array(val_Y)),  shuffle=False, callbacks=[es] ) 
+
+
+                    predicted_labels = gridmodel.predict([np.array(val_drugs), np.array(val_prots) ])
+                    loss, rperf2 = gridmodel.evaluate(([np.array(val_drugs),np.array(val_prots) ]), np.array(val_Y), verbose=0)
+                    rperf = prfmeasure(val_Y, predicted_labels)
+                    rperf = rperf[0]
+
+
+                    logging("P1 = %d,  P2 = %d, P3 = %d, Fold = %d, CI-i = %f, CI-ii = %f, MSE = %f" % 
+                    (param1ind, param2ind, param3ind, foldind, rperf, rperf2, loss), FLAGS)
+
+                    plotLoss(gridres, param1ind, param2ind, param3ind, foldind)
+
+                    all_predictions[pointer][foldind] =rperf #TODO FOR EACH VAL SET allpredictions[pointer][foldind]
+                    all_losses[pointer][foldind]= loss
+
+                    pointer +=1
+
+    bestperf = -float('Inf')
+    bestpointer = None
+
+
+    best_param_list = []
+    ##Take average according to folds, then chooose best params
+    pointer = 0
+    for param1ind in range(len(paramset1)):
+            for param2ind in range(len(paramset2)):
+                for param3ind in range(len(paramset3)):
+                
+                    avgperf = 0.
+                    for foldind in range(len(val_sets)):
+                        foldperf = all_predictions[pointer][foldind]
+                        avgperf += foldperf
+                    avgperf /= len(val_sets)
+                    #print(epoch, batchsz, avgperf)
+                    if avgperf > bestperf:
+                        bestperf = avgperf
+                        bestpointer = pointer
+                        best_param_list = [param1ind, param2ind, param3ind]
+
+                    pointer +=1
+        
+    return  bestpointer, best_param_list, bestperf, all_predictions, all_losses
+
+
+
+def cindex_score(y_true, y_pred):
+
+    g = tf.subtract(tf.expand_dims(y_pred, -1), y_pred)
+    g = tf.cast(g == 0.0, tf.float32) * 0.5 + tf.cast(g > 0.0, tf.float32)
+
+    f = tf.subtract(tf.expand_dims(y_true, -1), y_true) > 0.0
+    f = tf.matrix_band_part(tf.cast(f, tf.float32), -1, 0)
+
+    g = tf.reduce_sum(tf.multiply(g, f))
+    f = tf.reduce_sum(f)
+
+    return tf.where(tf.equal(g, 0), 0.0, g/f) #select
+
+
+   
+def plotLoss(history, batchind, epochind, param3ind, foldind):
+
+    figname = "b"+str(batchind) + "_e" + str(epochind) + "_" + str(param3ind) + "_"  + str( foldind) + "_" + str(time.time()) 
+    plt.figure()
+    plt.plot(history.history['loss'])
+    plt.plot(history.history['val_loss'])
+    plt.title('model loss')
+    plt.ylabel('loss')
+    plt.xlabel('epoch')
+	#plt.legend(['trainloss', 'valloss', 'cindex', 'valcindex'], loc='upper left')
+    plt.legend(['trainloss', 'valloss'], loc='upper left')
+    plt.savefig("figures/"+figname +".png" , dpi=None, facecolor='w', edgecolor='w', orientation='portrait', 
+                    papertype=None, format=None,transparent=False, bbox_inches=None, pad_inches=0.1,frameon=None)
+    plt.close()
+
+
+    ## PLOT CINDEX
+    plt.figure()
+    plt.title('model concordance index')
+    plt.ylabel('cindex')
+    plt.xlabel('epoch')
+    plt.plot(history.history['cindex_score'])
+    plt.plot(history.history['val_cindex_score'])
+    plt.legend(['traincindex', 'valcindex'], loc='upper left')
+    plt.savefig("figures/"+figname + "_acc.png" , dpi=None, facecolor='w', edgecolor='w', orientation='portrait', 
+                            papertype=None, format=None,transparent=False, bbox_inches=None, pad_inches=0.1,frameon=None)
+    plt.close()
+
+
+
+def prepare_interaction_pairs(XD, XT,  Y, rows, cols):
+    drugs = []
+    targets = []
+    targetscls = []
+    affinity=[] 
+        
+    for pair_ind in range(len(rows)):
+        drug = XD[rows[pair_ind]]
+        drugs.append(drug)
+
+        target=XT[cols[pair_ind]]
+        targets.append(target)
+
+        affinity.append(Y[rows[pair_ind],cols[pair_ind]])
+
+    drug_data = np.stack(drugs)
+    target_data = np.stack(targets)
+
+    return drug_data,target_data,  affinity
+
+
+       
+def experiment(FLAGS, perfmeasure, deepmethod, foldcount=6): #5-fold cross validation + test
+
+    #Input
+    #XD: [drugs, features] sized array (features may also be similarities with other drugs
+    #XT: [targets, features] sized array (features may also be similarities with other targets
+    #Y: interaction values, can be real values or binary (+1, -1), insert value float("nan") for unknown entries
+    #perfmeasure: function that takes as input a list of correct and predicted outputs, and returns performance
+    #higher values should be better, so if using error measures use instead e.g. the inverse -error(Y, P)
+    #foldcount: number of cross-validation folds for settings 1-3, setting 4 always runs 3x3 cross-validation
+
+
+    dataset = DataSet( fpath = FLAGS.dataset_path, ### BUNU ARGS DA GUNCELLE
+                      setting_no = FLAGS.problem_type, ##BUNU ARGS A EKLE
+                      seqlen = FLAGS.max_seq_len,
+                      smilen = FLAGS.max_smi_len,
+                      need_shuffle = False )
+    # set character set size
+    FLAGS.charseqset_size = dataset.charseqset_size 
+    FLAGS.charsmiset_size = dataset.charsmiset_size 
+
+    XD, XT, Y = dataset.parse_data(FLAGS)
+
+    XD = np.asarray(XD)
+    XT = np.asarray(XT)
+    Y = np.asarray(Y)
+
+    drugcount = XD.shape[0]
+    print(drugcount)
+    targetcount = XT.shape[0]
+    print(targetcount)
+
+    FLAGS.drug_count = drugcount
+    FLAGS.target_count = targetcount
+
+    label_row_inds, label_col_inds = np.where(np.isnan(Y)==False)  #basically finds the point address of affinity [x,y]
+
+    if not os.path.exists(figdir):
+        os.makedirs(figdir)
+
+    print(FLAGS.log_dir)
+    S1_avgperf, S1_avgloss, S1_teststd = nfold_1_2_3_setting_sample(XD, XT, Y, label_row_inds, label_col_inds,
+                                                                     perfmeasure, deepmethod, FLAGS, dataset)
+
+    logging("Setting " + str(FLAGS.problem_type), FLAGS)
+    logging("avg_perf = %.5f,  avg_mse = %.5f, std = %.5f" % 
+            (S1_avgperf, S1_avgloss, S1_teststd), FLAGS)
+
+
+
+
+def run_regression( FLAGS ): 
+
+    perfmeasure = get_cindex
+    deepmethod = build_combined_categorical
+
+    experiment(FLAGS, perfmeasure, deepmethod)
+
+
+
+
+if __name__=="__main__":
+    
+    FLAGS = argparser()
+
+    FLAGS.num_windows = [32]
+    FLAGS.seq_window_lengths = [8, 12]
+    FLAGS.smi_window_lengths = [4, 8]
+    FLAGS.batch_size = 256
+    FLAGS.num_epoch = 100
+    FLAGS.max_seq_len = 1000
+    FLAGS.max_smi_len = 100
+    FLAGS.dataset_path = 'data/kiba/'
+    FLAGS.problem_type = 1
+    FLAGS.log_dir = "logs"
+    
+    FLAGS.log_dir = FLAGS.log_dir + str(time.time()) + "/"
+
+    
+    if not os.path.exists(FLAGS.log_dir):
+        os.makedirs(FLAGS.log_dir)
+
+    logging(str(FLAGS), FLAGS)
+    run_regression( FLAGS )

baselines/DeepDTA/source/train.sh

+python run_baseline.py 
\ No newline at end of file

MolTrans @ 47ac16b8

-Subproject commit 47ac16b8c158b080ba6cdaec74cd7aa9c1332b73