def InputROOT_OutputTXT(infileROOT, ModelName):
inputFile = read_file_name_root(infileROOT)
ROOT_input = inputFile[2]
Text_output = inputFile[0] + ".txt"
# data = TFile.Open('/Users/leejunho/Desktop/git/PKUHEP/DNN/SSWW_split_input/result/for_DNN/TTTL_LL_230M_comparable.root')
data = TFile.Open(ROOT_input)
tree = data.Get('tree')
####################################### Input DATA Sets !!!!!
lep1pt_ = tree2array(tree, branches='lep1pt')
lep1eta_ = tree2array(tree, branches='lep1eta')
lep1phi_ = tree2array(tree, branches='lep1phi')
lep2pt_ = tree2array(tree, branches='lep2pt')
lep2eta_ = tree2array(tree, branches='lep2eta')
lep2phi_ = tree2array(tree, branches='lep2phi')
jet1pt_ = tree2array(tree, branches='jet1pt')
jet1eta_ = tree2array(tree, branches='jet1eta')
jet1phi_ = tree2array(tree, branches='jet1phi')
jet1M_ = tree2array(tree, branches='jet1M')
jet2pt_ = tree2array(tree, branches='jet2pt')
jet2eta_ = tree2array(tree, branches='jet2eta')
jet2phi_ = tree2array(tree, branches='jet2phi')
jet2M_ = tree2array(tree, branches='jet2M')
MET_ = tree2array(tree, branches='MET')
#lep1PID_ = tree2array(tree, branches='lep1PID')
#lep2PID_ = tree2array(tree, branches='lep2PID')
#Mjj_ = tree2array(tree, branches='Mjj')
#dr_ll_jj_ = tree2array(tree, branches='dr_ll_jj')
#dphijj_ = tree2array(tree, branches='dphijj')
#zeppen_lep1_ = tree2array(tree, branches='zeppen_lep1')
#zeppen_lep2_ = tree2array(tree, branches='zeppen_lep2')
METphi_ = tree2array(tree, branches='METphi')
#detajj_ = tree2array(tree, branches='detajj')
#Mll_ = tree2array(tree, branches='Mll')
#RpT_ = tree2array(tree, branches='RpT')
###############################################################################################################
##################################### Target DATA !!!!!
LL_Helicity_ = tree2array(tree, branches='LL_Helicity')
TTTL_Helicity_ = tree2array(tree, branches='TTTL_Helicity')
#TL_Helicity_ = tree2array(tree, branches='TL_Helicity')
#TT_Helicity_ = tree2array(tree, branches='TT_Helicity')
###############################################################################################################
ENTRY = LL_Helicity_.size
print "ENTRY :", ENTRY
lep1pt = np.zeros(ENTRY)
lep1eta = np.zeros(ENTRY)
lep1phi = np.zeros(ENTRY)
lep2pt = np.zeros(ENTRY)
lep2eta = np.zeros(ENTRY)
lep2phi = np.zeros(ENTRY)
jet1pt = np.zeros(ENTRY)
jet1eta = np.zeros(ENTRY)
jet1phi = np.zeros(ENTRY)
jet1M = np.zeros(ENTRY)
jet2pt = np.zeros(ENTRY)
jet2eta = np.zeros(ENTRY)
jet2phi = np.zeros(ENTRY)
jet2M = np.zeros(ENTRY)
MET = np.zeros(ENTRY)
#lep1PID = np.zeros(ENTRY)
#lep2PID = np.zeros(ENTRY)
#Mjj = np.zeros(ENTRY)
#dr_ll_jj = np.zeros(ENTRY)
#dphijj = np.zeros(ENTRY)
#zeppen_lep1 = np.zeros(ENTRY)
#zeppen_lep2 = np.zeros(ENTRY)
METphi = np.zeros(ENTRY)
#detajj = np.zeros(ENTRY)
#Mll = np.zeros(ENTRY)
#RpT = np.zeros(ENTRY)
LL_Helicity = np.zeros(ENTRY)
TTTL_Helicity = np.zeros(ENTRY)
#TL_Helicity = np.zeros(ENTRY)
#TT_Helicity = np.zeros(ENTRY)
###############################################################################################################
for j1 in range(ENTRY):
lep1pt[j1] = lep1pt_[j1]
lep1eta[j1] = lep1eta_[j1]
lep1phi[j1] = lep1phi_[j1]
lep2pt[j1] = lep2pt_[j1]
lep2eta[j1] = lep2eta_[j1]
lep2phi[j1] = lep2phi_[j1]
jet1pt[j1] = jet1pt_[j1]
jet1eta[j1] = jet1eta_[j1]
jet1phi[j1] = jet1phi_[j1]
jet1M[j1] = jet1M_[j1]
jet2pt[j1] = jet2pt_[j1]
jet2eta[j1] = jet2eta_[j1]
jet2phi[j1] = jet2phi_[j1]
jet2M[j1] = jet2M_[j1]
MET[j1] = MET_[j1]
#lep1PID[j1] = lep1PID_[j1]
#lep2PID[j1] = lep2PID_[j1]
#Mjj[j1] = Mjj_[j1]
#dr_ll_jj[j1] = dr_ll_jj_[j1]
#dphijj[j1] = dphijj_[j1]
#zeppen_lep1[j1] = zeppen_lep1_[j1]
#zeppen_lep2[j1] = zeppen_lep2_[j1]
METphi[j1] = METphi_[j1]
#detajj[j1] = detajj_[j1]
#Mll[j1] = Mll_[j1]
#RpT[j1] = RpT_[j1]
LL_Helicity[j1] = LL_Helicity_[j1]
TTTL_Helicity[j1] = TTTL_Helicity_[j1]
#TL_Helicity[j1] = TL_Helicity_[j1]
#TT_Helicity[j1] = TT_Helicity_[j1]
#ARRAY = np.stack((lep1pt, lep1eta, lep1phi, lep2pt, lep2eta, lep2phi, jet1pt, jet1eta, jet1phi, jet1M, jet2pt, jet2eta, jet2phi, jet2M, MET, lep1PID, lep2PID, Mjj, dr_ll_jj, dphijj, zeppen_lep1, zeppen_lep2, METphi, detajj, Mll, RpT))
#ARRAY = np.stack((lep1pt, lep1eta, lep2pt, lep2eta, jet1pt, jet1eta, jet2pt, jet2eta, MET, dr_ll_jj, dphijj, detajj, Mll))
ARRAY = np.stack(
(lep1pt, lep1eta, lep1phi, lep2pt, lep2eta, lep2phi, jet1pt, jet1eta,
jet1phi, jet1M, jet2pt, jet2eta, jet2phi, jet2M, MET, METphi))
TARGET = np.stack((LL_Helicity, TTTL_Helicity))
#TARGET = np.stack((LL_Helicity, TL_Helicity, TT_Helicity))
ARRAY = ARRAY.T
TARGET = TARGET.T
##'''
X_part = ARRAY[:]
Y_part = TARGET[:]
#N_validation = ARRAY.shape[0]-(N_train)
print(X_part.shape, "x_train")
print(Y_part.shape)
#model = DNN(n_in=26, n_hiddens=[150], n_out=2) ##FIXME TODO
model = DNN(n_in=16,
n_hiddens=[150, 150, 150, 150, 150, 150, 150, 150, 150, 150],
n_out=2) ##FIXME TODO
print("Error test1")
model.fit_classify_model_read(ModelName=ModelName)
print("Error test2")
accuracy = model.evaluate(X_part, Y_part)
print('accuracy:', accuracy)
np.set_printoptions(threshold='nan')
LL_TTTL_prob_tuple = model.Indicate_classified_LL_TTTL(X_part, Y_part)
OF = open(Text_output, "w+")
print(len(LL_TTTL_prob_tuple[2]))
for i in range(len(LL_TTTL_prob_tuple[2])):
if i == 0: OF.write("%s\n" % "LL TTTL")
LL_n_TTTL = str(LL_TTTL_prob_tuple[2][i]).replace("[", "")
LL_n_TTTL = LL_n_TTTL.replace("]", "")
OF.write("%s" % LL_n_TTTL)
#LL_p, TTTL_p = LL_n_TTTL.split(); LL_p = float(LL_p); TTTL_p = float(TTTL_p)
#print(LL_n_TTTL, LL_p, TTTL_p)
OF.write("\n")
OF.close()
##'''
'''
X_train = ARRAY[:]
Y_train = TARGET[:]
N_validation = ARRAY.shape[0]-(N_train)
X_train, X_test, Y_train, Y_test = train_test_split(X_train, Y_train, train_size=N_train)
X_train, X_validation, Y_train, Y_validation = train_test_split(X_train, Y_train, test_size=N_validation)
print(X_train.shape,"x_train");print(X_validation.shape,"x_validation");print(Y_train.shape);print(Y_validation.shape)
model = DNN(n_in=26, n_hiddens=[150,150,150], n_out=2)
model.fit_classify_model_read(ModelName=ModelName)
accuracy = model.evaluate(X_test, Y_test)
print('accuracy:', accuracy)
np.set_printoptions(threshold='nan')
LL_TTTL_prob_tuple = model.Indicate_classified_LL_TTTL(X_test,Y_test)
'''
'''
def InputROOT_OutputTXT(infileROOT,ModelName):
inputFile = read_file_name_root(infileROOT)
ROOT_input = inputFile[2]
ROOT_Estimated_LL = inputFile[3] + "Estimated_LL.root"
ROOT_Estimated_TTTL = inputFile[3] + "Estimated_TTTL.root"
# ROOT_Estimated_TTTL = inputFile[3] + inputFile[0] + "_Esti_TTTL.root"
print(ROOT_Estimated_LL, ROOT_Estimated_TTTL)
data = TFile.Open(ROOT_input)
tree = data.Get('tree')
####################################### Input DATA Sets !!!!!
lep1pt_ = tree2array(tree, branches='lep1pt')
lep1eta_ = tree2array(tree, branches='lep1eta')
#lep1phi_ = tree2array(tree, branches='lep1phi')
lep2pt_ = tree2array(tree, branches='lep2pt')
lep2eta_ = tree2array(tree, branches='lep2eta')
#lep2phi_ = tree2array(tree, branches='lep2phi')
jet1pt_ = tree2array(tree, branches='jet1pt')
jet1eta_ = tree2array(tree, branches='jet1eta')
#jet1phi_ = tree2array(tree, branches='jet1phi')
#jet1M_ = tree2array(tree, branches='jet1M')
jet2pt_ = tree2array(tree, branches='jet2pt')
jet2eta_ = tree2array(tree, branches='jet2eta')
#jet2phi_ = tree2array(tree, branches='jet2phi')
#jet2M_ = tree2array(tree, branches='jet2M')
MET_ = tree2array(tree, branches='MET')
#lep1PID_ = tree2array(tree, branches='lep1PID')
#lep2PID_ = tree2array(tree, branches='lep2PID')
#Mjj_ = tree2array(tree, branches='Mjj')
dr_ll_jj_ = tree2array(tree, branches='dr_ll_jj')
dphijj_ = tree2array(tree, branches='dphijj')
#zeppen_lep1_ = tree2array(tree, branches='zeppen_lep1')
#zeppen_lep2_ = tree2array(tree, branches='zeppen_lep2')
#METphi_ = tree2array(tree, branches='METphi')
detajj_ = tree2array(tree, branches='detajj')
Mll_ = tree2array(tree, branches='Mll')
#RpT_ = tree2array(tree, branches='RpT')
###############################################################################################################
##################################### Target DATA !!!!!
LL_Helicity_ = tree2array(tree, branches='LL_Helicity')
TTTL_Helicity_ = tree2array(tree, branches='TTTL_Helicity')
#TL_Helicity_ = tree2array(tree, branches='TL_Helicity')
#TT_Helicity_ = tree2array(tree, branches='TT_Helicity')
###############################################################################################################
ENTRY = LL_Helicity_.size
print "ENTRY :", ENTRY
lep1pt = np.zeros(ENTRY)
lep1eta = np.zeros(ENTRY)
#lep1phi = np.zeros(ENTRY)
lep2pt = np.zeros(ENTRY)
lep2eta = np.zeros(ENTRY)
#lep2phi = np.zeros(ENTRY)
jet1pt = np.zeros(ENTRY)
jet1eta = np.zeros(ENTRY)
#jet1phi = np.zeros(ENTRY)
#jet1M = np.zeros(ENTRY)
jet2pt = np.zeros(ENTRY)
jet2eta = np.zeros(ENTRY)
#jet2phi = np.zeros(ENTRY)
#jet2M = np.zeros(ENTRY)
MET = np.zeros(ENTRY)
#lep1PID = np.zeros(ENTRY)
#lep2PID = np.zeros(ENTRY)
#Mjj = np.zeros(ENTRY)
dr_ll_jj = np.zeros(ENTRY)
dphijj = np.zeros(ENTRY)
#zeppen_lep1 = np.zeros(ENTRY)
#zeppen_lep2 = np.zeros(ENTRY)
#METphi = np.zeros(ENTRY)
detajj = np.zeros(ENTRY)
Mll = np.zeros(ENTRY)
#RpT = np.zeros(ENTRY)
LL_Helicity = np.zeros(ENTRY)
TTTL_Helicity = np.zeros(ENTRY)
#TL_Helicity = np.zeros(ENTRY)
#TT_Helicity = np.zeros(ENTRY)
###############################################################################################################
for j1 in range(ENTRY):
lep1pt[j1] = lep1pt_[j1]
lep1eta[j1] = lep1eta_[j1]
#lep1phi[j1] = lep1phi_[j1]
lep2pt[j1] = lep2pt_[j1]
lep2eta[j1] = lep2eta_[j1]
#lep2phi[j1] = lep2phi_[j1]
jet1pt[j1] = jet1pt_[j1]
jet1eta[j1] = jet1eta_[j1]
#jet1phi[j1] = jet1phi_[j1]
#jet1M[j1] = jet1M_[j1]
jet2pt[j1] = jet2pt_[j1]
jet2eta[j1] = jet2eta_[j1]
#jet2phi[j1] = jet2phi_[j1]
#jet2M[j1] = jet2M_[j1]
MET[j1] = MET_[j1]
#lep1PID[j1] = lep1PID_[j1]
#lep2PID[j1] = lep2PID_[j1]
#Mjj[j1] = Mjj_[j1]
dr_ll_jj[j1] = dr_ll_jj_[j1]
dphijj[j1] = dphijj_[j1]
#zeppen_lep1[j1] = zeppen_lep1_[j1]
#zeppen_lep2[j1] = zeppen_lep2_[j1]
#METphi[j1] = METphi_[j1]
detajj[j1] = detajj_[j1]
Mll[j1] = Mll_[j1]
#RpT[j1] = RpT_[j1]
LL_Helicity[j1] = LL_Helicity_[j1]
TTTL_Helicity[j1] = TTTL_Helicity_[j1]
#TL_Helicity[j1] = TL_Helicity_[j1]
#TT_Helicity[j1] = TT_Helicity_[j1]
#ARRAY = np.stack((lep1pt, lep1eta, lep1phi, lep2pt, lep2eta, lep2phi, jet1pt, jet1eta, jet1phi, jet1M, jet2pt, jet2eta, jet2phi, jet2M, MET, lep1PID, lep2PID, Mjj, dr_ll_jj, dphijj, zeppen_lep1, zeppen_lep2, METphi, detajj, Mll, RpT))
ARRAY = np.stack((lep1pt, lep1eta, lep2pt, lep2eta, jet1pt, jet1eta, jet2pt, jet2eta, MET, dr_ll_jj, dphijj, detajj, Mll))
TARGET = np.stack((LL_Helicity, TTTL_Helicity))
#TARGET = np.stack((LL_Helicity, TL_Helicity, TT_Helicity))
ARRAY = ARRAY.T
TARGET = TARGET.T
##'''
X_part = ARRAY[:]
Y_part = TARGET[:]
#N_validation = ARRAY.shape[0]-(N_train)
print(X_part.shape,"x_train"); print(Y_part.shape)
#model = DNN(n_in=26, n_hiddens=[150], n_out=2) ##FIXME TODO
model = DNN(n_in=13, n_hiddens=[150,150,150,150,150,150,150], n_out=2) ##FIXME TODO
model.fit_classify_model_read(ModelName=ModelName)
accuracy = model.evaluate(X_part, Y_part)
print('accuracy:', accuracy)
np.set_printoptions(threshold='nan')
LL_TTTL_prob_tuple = model.Indicate_classified_LL_TTTL(X_part, Y_part)
print(len(LL_TTTL_prob_tuple[2]))
#List = [];
List_LL = []; List_TTTL = []
for i in range(len(LL_TTTL_prob_tuple[2])):
LL_n_TTTL = str(LL_TTTL_prob_tuple[2][i]).replace("[","")
LL_n_TTTL = LL_n_TTTL.replace("]","")
XY_part = np.append(X_part[i],Y_part[i])
XY_part = tuple(XY_part)
#List.append(XY_part)
LL_p, TTTL_p = LL_n_TTTL.split(); LL_p = float(LL_p); TTTL_p = float(TTTL_p)
#print(LL_n_TTTL, LL_p, TTTL_p)
if(LL_p>0.5):
List_LL.append(XY_part)
elif(TTTL_p>0.5):
List_TTTL.append(XY_part)
else:
print("ERROR on LL, TTTL proportion!")
break
nplist_LL = np.array(List_LL, dtype=[('lep1pt',np.float32),('lep1eta',np.float32),('lep2pt',np.float32),('lep2eta',np.float32),
('jet1pt',np.float32),('jet1eta',np.float32),('jet2pt',np.float32),('jet2eta',np.float32),
('MET',np.float32),('dr_ll_jj',np.float32),('dphijj',np.float32),
('detajj',np.float32),('Mll',np.float32), ('LL_Helicity',np.float32), ('TTTL_Helicity',np.float32)] )
nplist_TTTL = np.array(List_TTTL, dtype=[('lep1pt',np.float32),('lep1eta',np.float32),('lep2pt',np.float32),('lep2eta',np.float32),
('jet1pt',np.float32),('jet1eta',np.float32),('jet2pt',np.float32),('jet2eta',np.float32),
('MET',np.float32),('dr_ll_jj',np.float32),('dphijj',np.float32),
('detajj',np.float32),('Mll',np.float32), ('LL_Helicity',np.float32), ('TTTL_Helicity',np.float32)] )
LL_ROOT = TFile(ROOT_Estimated_LL,"RECREATE")
tree_LL = array2tree(nplist_LL)
tree_LL.Write()
LL_ROOT.Close()
TTTL_ROOT = TFile(ROOT_Estimated_TTTL,"RECREATE")
tree_TTTL = array2tree(nplist_TTTL)
tree_TTTL.Write()
TTTL_ROOT.Close()
return [ROOT_Estimated_LL,ROOT_Estimated_TTTL]
tree_train_LL = array2tree(TRAIN_nplist_LL)
tree_train_LL.Write()
Train_ROOT_LL.Close()
Train_ROOT_TTTL = TFile(ROOT_filename_TTTL, "RECREATE")
tree_train_TTTL = array2tree(TRAIN_nplist_TTTL)
tree_train_TTTL.Write()
Train_ROOT_TTTL.Close()
del Train_List
del TRAIN_nplist
## </SAVE TRAIN ROOT>
epoch_txt_loca = "tens_model_class/" + Model_name
model.fit_classify(X_train,
Y_train,
X_validation,
Y_validation,
epochs=epochs,
batch_size=batch_size,
p_keep=0.5,
earlyStop=earlyStop,
model_name=model_name,
epoch_txt_loca=epoch_txt_loca)
accuracy = model.evaluate(X_test, Y_test)
print('accuracy:', accuracy)
plot_name = "tens_model_class/" + Model_name + "/" + "SSWW_classification_TTTL-LL_comp.pdf"
model.Plot_acc_loss(plot_name=plot_name)
Input_txt = epoch_txt_loca + "/Loss_epoch.txt"
Plotting(text_infile=Input_txt)
TARGET = TARGET.T
X_part = ARRAY[:]
Y_part = TARGET[:]
N_validation = ARRAY.shape[0] - (N_train)
#X_train, X_test, Y_train, Y_test = train_test_split(X_train, Y_train, train_size=N_train)
#X_train, X_validation, Y_train, Y_validation = train_test_split(X_train, Y_train, test_size=N_validation)
#print(X_train.shape,"x_train");print(X_validation.shape,"x_validation");print(Y_train.shape);print(Y_validation.shape)
print(X_part.shape, "x_train")
print(Y_part.shape)
model = DNN(n_in=26, n_hiddens=[150, 150, 150], n_out=2)
model.fit_classify_model_read(ModelName=ModelName)
#accuracy = model.evaluate(X_test, Y_test)
accuracy = model.evaluate(X_part, Y_part)
print('accuracy:', accuracy)
np.set_printoptions(threshold='nan')
LL_TTTL_prob_tuple = model.Indicate_classified_LL_TTTL(X_part, Y_part)
#print(LL_TTTL_prob.shape)
#print(X_part.shape)
#print(X_part[0])
#LL_nplist = np.zeros(LL_TTTL_prob_tuple[0])
#TTTL_nplist = np.zeros(LL_TTTL_prob_tuple[1])
#print(LL_nplist.shape)
#print(TTTL_nplist.shape)
LL_list = []
TTTL_list = []