def compare_sharptni_tnet_cdc(threshold):
data_dir = 'CDC/'
folders = next(os.walk(data_dir))[1]
folders.sort()
F1_file = open('results/sharptni/cdc.bestTree.sharptni.sankoff_sample.tnet.new.rand.mod.th.'+str(threshold)+'.csv', 'w+')
F1_file.write('dataset,sharp_prec,sharp_rec,sharp_f1,tnet_prec,tnet_rec,tnet_f1\n')
for folder in folders:
print('inside folder: ',folder)
F1 = []
sample_list = next(os.walk(data_dir + folder + '/sharptni_output'))[2]
sharptni_file = [idx for idx in sample_list if idx.startswith('sample_sankoff_summary')]
sharptni_file = sharptni_file[0]
th2 = int(sharptni_file.split('.')[1])
th2 = round(th2 * (threshold / 100))
print(th2)
real = set(cdc.get_true_transmission_edges(folder))
sharp = set(ge.get_mul_tnet_edges(data_dir + folder + '/sharptni_output/' + sharptni_file, th2))
tnet = set(ge.get_mul_tnet_edges(data_dir + folder + '/tnet_new_mod_rand_bootstrap/25.tnet', threshold))
F1.extend(get_prec_rec_f1(real, sharp))
F1.extend(get_prec_rec_f1(real, tnet))
F1_file.write('{},{},{},{},{},{},{}\n'.format(folder,F1[0],F1[1],F1[2],F1[3],F1[4],F1[5]))
F1_file.close()
def compare_favites_best_tree_sharptni_tnet_new_tnet_bias_directed(sample_th):
F1_file = open('results/single_tree_sharptni/favites.best_tree.sharptni_min_coinf.tnet_new.tnet_bias.sample_th.' + str(sample_th) + '.csv', 'w+')
F1_file.write('dataset,sharp_prec,sharp_rec,sharp_f1,tnet_prec,tnet_rec,tnet_f1,tnet_bias_prec,tnet_bias_rec,tnet_bias_f1\n')
folders = next(os.walk('outputs/'))[1]
for folder in folders:
print('inside folder:', folder)
F1 = []
sample_list = next(os.walk('outputs/' + folder + '/sharptni_single'))[2]
sharptni_file = [idx for idx in sample_list if idx.startswith('bestTree_sankoff_min_coinfection.100_sample')]
sharptni_file = sharptni_file[0]
sample_num = int(sharptni_file.split('.')[2])
sharp_th = math.ceil(sample_num * (sample_th / 100))
print(sample_num, sharp_th)
real = set(ge.get_real_edges('dataset/' + folder + '/transmission_network.txt'))
sharptni = set(ge.get_mul_tnet_edges('outputs/' + folder + '/sharptni_single/' + sharptni_file, sharp_th))
tnet = set(ge.get_mul_tnet_edges('outputs/' + folder + '/tnet_best_tree/bestTree.100.tnet_new', sample_th))
tnet_bias = set(ge.get_mul_tnet_edges('outputs/' + folder + '/tnet_best_tree/bestTree.100.tnet_new_with_bias', sample_th))
F1.extend(get_prec_rec_f1(real, sharptni))
F1.extend(get_prec_rec_f1(real, tnet))
F1.extend(get_prec_rec_f1(real, tnet_bias))
F1_file.write('{},{},{},{},{},{},{},{},{},{}\n'.format(folder,F1[0],F1[1],F1[2],F1[3],F1[4],F1[5],F1[6],F1[7],F1[8]))
F1_file.close()
def create_directed_tnet_bootstrap_summary(tree_folder, threshold):
data_dir = 'covid_19/NCBI/'
edge_dict = {}
bootstrap_folder = data_dir + tree_folder
output_folder = data_dir + '/tnet_output_complete/'
if not os.path.exists(output_folder):
os.mkdir(output_folder)
if not os.path.exists(output_folder + 'bootstrap_tnet_bias_100_th_' +
str(threshold) + '_summary.csv'):
result = open(
output_folder + 'bootstrap_tnet_bias_100_th_' + str(threshold) +
'_summary.csv', 'w+')
file_list = next(os.walk(bootstrap_folder))[2]
for file in file_list:
tnet_file = bootstrap_folder + '/' + file
tnet_edges = ge.get_mul_tnet_edges(tnet_file, threshold)
for edge in tnet_edges:
if edge in edge_dict:
edge_dict[edge] += 1
else:
edge_dict[edge] = 1
edge_dict = dict(
sorted(edge_dict.items(), key=operator.itemgetter(1),
reverse=True))
for x, y in edge_dict.items():
result.write('{},{}\n'.format(x, y))
def compare_sharptni_best_tree():
data_dir = 'outputs/'
folders = next(os.walk(data_dir))[1]
folders.sort()
thresholds = [10, 20, 30, 40, 50, 60, 70, 80, 90, 100]
F1_file = open('results/sharptni/best_tree.recall.sample_sankoff.csv', 'w+')
F1_file.write('dataset,10,20,30,40,50,60,70,80,90,100\n')
for folder in folders:
print('inside folder: ',folder)
real = set(ge.get_real_edges('dataset/' + folder + '/transmission_network.txt'))
sample_list = next(os.walk(data_dir + folder + '/sharptni'))[2]
sharptni_file = [idx for idx in sample_list if idx.startswith('sample_sankoff_summary')]
sharptni_file = sharptni_file[0]
sample_num = int(sharptni_file.split('.')[1])
print(sample_num)
F1 = []
for th in thresholds:
thr = round(sample_num * (th / 100))
tnet = set(ge.get_mul_tnet_edges(data_dir + folder + '/sharptni/' + sharptni_file, thr))
temp = get_prec_rec_f1(real, tnet)
F1.append(temp[1])
F1_file.write('{},{},{},{},{},{},{},{},{},{},{}\n'.format(folder,F1[0],F1[1],F1[2],F1[3],F1[4],F1[5]
,F1[6],F1[7],F1[8],F1[9]))
def create_cdc_tnet_summary_directed(threshold):
for outbreak in known_outbreaks:
print('Inside', outbreak)
input_folder = 'CDC/' + outbreak + '/tnet_new_bootstrap'
output_folder = 'CDC/' + outbreak + '/tnet_new_bootstrap_summary_directed'
if not os.path.exists(output_folder):
os.mkdir(output_folder)
edge_dict = {}
result = open(
output_folder + '/tnet_new_bootstrap' + '_th_' + str(threshold) +
'_summary.csv', 'w+')
file_list = next(os.walk(input_folder))[2]
for file in file_list:
tnet_file = input_folder + '/' + file
tnet_edges = ge.get_mul_tnet_edges(tnet_file, threshold)
for edge in tnet_edges:
if edge in edge_dict:
edge_dict[edge] += 1
else:
edge_dict[edge] = 1
edge_dict = dict(
sorted(edge_dict.items(), key=operator.itemgetter(1),
reverse=True))
for x, y in edge_dict.items():
result.write('{},{}\n'.format(x, y))
def compare_tnet_cdc_single_tree():
F1_file = open('results/cdc_single_tree_tnet/single_tree.f1.tnet.new.with.min.csv', 'w+')
F1_file.write('dataset,single,10,20,30,40,50,60,70,80,90,100\n')
thresholds = [10, 20, 30, 40, 50, 60, 70, 80, 90, 100]
for outbreak in cdc.known_outbreaks:
real = set(cdc.get_true_transmission_edges(outbreak))
tnet_single = set(ge.get_mul_tnet_edges('CDC/' + outbreak + '/tnet_single_tree/single_tree.1.tnet_new_min', 0))
single_run = get_prec_rec_f1(real, tnet_single)[2]
F1 = []
for th in thresholds:
tnet = set(ge.get_mul_tnet_edges('CDC/' + outbreak + '/tnet_single_tree/single_tree.100.tnet_new_min', th))
temp = get_prec_rec_f1(real, tnet)
F1.append(temp[2])
F1_file.write('{},{},{},{},{},{},{},{},{},{},{},{}\n'.format(outbreak,single_run,F1[0],F1[1],F1[2],F1[3],F1[4],F1[5]
,F1[6],F1[7],F1[8],F1[9]))
def compare_favites_sharptni_phyloscanner_tnet_new_tnet_bias_single_tree_single_run():
F1_file = open('results/single_tree_sharptni/favites.phyloscanner.sharptni.min.coinf.tnet.new.tnet.bias.single_tree.single_run.csv', 'w+')
F1_file.write('dataset,phylo_prec,phylo_rec,phylo_f1,sharp_prec,sharp_rec,sharp_f1,tnet_prec,tnet_rec,tnet_f1,tnet_bias_prec,tnet_bias_rec,tnet_bias_f1\n')
phylo_dir = '/home/saurav/research/FAVITES_compare_TNet_v2/outputs/'
folders = next(os.walk('outputs/'))[1]
for folder in folders:
print('inside folder:', folder)
F1 = []
real = set(ge.get_real_edges('dataset/' + folder + '/transmission_network.txt'))
phylo = set(ge.get_phyloscanner_single_tree_edges(phylo_dir + folder + '/phyloscanner_best_tree/favites_collapsedTree.csv'))
sharptni = set(ge.get_mul_tnet_edges('outputs/' + folder + '/sharptni_single/bestTree_sankoff_min_coinfection.1', 1))
tnet = set(ge.get_mul_tnet_edges('outputs/' + folder + '/tnet_best_tree/bestTree.1.tnet_new', 1))
tnet_bias = set(ge.get_mul_tnet_edges('outputs/' + folder + '/tnet_best_tree/bestTree.1.tnet_new_with_bias', 1))
F1.extend(get_prec_rec_f1(real, phylo))
F1.extend(get_prec_rec_f1(real, sharptni))
F1.extend(get_prec_rec_f1(real, tnet))
F1.extend(get_prec_rec_f1(real, tnet_bias))
F1_file.write('{},{},{},{},{},{},{},{},{},{},{},{},{}\n'.format(folder,F1[0],F1[1],F1[2],F1[3],F1[4],F1[5],F1[6],F1[7],F1[8],F1[9],F1[10],F1[11]))
F1_file.close()
def compare_tnet_single_run():
data_dir = 'outputs/'
folders = next(os.walk(data_dir))[1]
folders.sort()
F1_file = open('results/single_tree_tnet/best_tree.tnet.new.50.csv', 'w+')
F1_file.write('dataset,precision,recall,f1\n')
for folder in folders:
print('inside folder: ',folder)
real = set(ge.get_real_edges('dataset/' + folder + '/transmission_network.txt'))
tnet_single = set(ge.get_mul_tnet_edges(data_dir + folder + '/tnet_best_tree/bestTree.100.tnet_new', 50))
F1 = get_prec_rec_f1(real, tnet_single)
F1_file.write('{},{},{},{}\n'.format(folder,F1[0],F1[1],F1[2]))
def compare_tnet_best_tree():
data_dir = 'outputs/'
folders = next(os.walk(data_dir))[1]
folders.sort()
thresholds = [10, 20, 30, 40, 50, 60, 70, 80, 90, 100]
F1_file = open('results/single_tree_tnet/best_tree.recall.tnet.new.csv', 'w+')
F1_file.write('dataset,single,10,20,30,40,50,60,70,80,90,100\n')
for folder in folders:
print('inside folder: ',folder)
real = set(ge.get_real_edges('dataset/' + folder + '/transmission_network.txt'))
tnet_single = set(ge.get_mul_tnet_edges(data_dir + folder + '/tnet_best_tree/bestTree.1.tnet_new', 0))
single_run = get_prec_rec_f1(real, tnet_single)[1]
F1 = []
for th in thresholds:
tnet = set(ge.get_mul_tnet_edges(data_dir + folder + '/tnet_best_tree/bestTree.100.tnet_new', th))
temp = get_prec_rec_f1(real, tnet)
F1.append(temp[1])
F1_file.write('{},{},{},{},{},{},{},{},{},{},{},{}\n'.format(folder,single_run,F1[0],F1[1],F1[2],F1[3],F1[4],F1[5]
,F1[6],F1[7],F1[8],F1[9]))
def compare_phyloscanner_tnet_best_tree(threshold):
data_dir = 'outputs/'
folders = next(os.walk(data_dir))[1]
folders.sort()
F1_file = open('results/best_tree.phyloscanner.tnet.new.th.'+str(threshold)+'.csv', 'w+')
F1_file.write('dataset,phylo_prec,phylo_rec,phylo_f1,tnet_prec,tnet_rec,tnet_f1\n')
for folder in folders:
print('inside folder: ',folder)
F1 = []
real = set(ge.get_real_edges('dataset/' + folder + '/transmission_network.txt'))
phylo = set(ge.get_phyloscanner_single_tree_edges(data_dir + folder + '/phyloscanner_best_tree/favites_collapsedTree.csv'))
tnet = set(ge.get_mul_tnet_edges(data_dir + folder + '/tnet_best_tree/bestTree.100.tnet_new', threshold))
F1.extend(get_prec_rec_f1(real, phylo))
F1.extend(get_prec_rec_f1(real, tnet))
F1_file.write('{},{},{},{},{},{},{}\n'.format(folder,F1[0],F1[1],F1[2],F1[3],F1[4],F1[5]))
F1_file.close()
def compare_tnet_best_tree():
data_dir = 'dataset/'
folders = next(os.walk(data_dir))[1]
folders.sort()
thresholds = [50, 60, 70, 80, 90, 100]
F1_file = open('results/old.prec.rec.f1.tnet.csv', 'w+')
F1_file.write('dataset,prec_50,rec_50,f1_50,prec_60,rec_60,f1_60,prec_70,rec_70,f1_70,prec_80,rec_80,f1_80,prec_90,rec_90,f1_90,prec_100,rec_100,f1_100\n')
for folder in folders:
print('inside folder: ',folder)
F1 = []
for th in thresholds:
real = set(ge.get_real_edges(data_dir + folder + '/transmission_network.txt'))
tnet = set(ge.get_mul_tnet_edges(data_dir + folder + '/tnet_old_100.tnet', th))
# tnet_new = set(ge.get_mul_tnet_edges(data_dir + folder + '/tnet_new_100.tnet', th))
temp = get_prec_rec_f1(real, tnet)
F1.extend(temp)
F1_file.write('{},{},{},{},{},{},{},{},{},{},{},{},{},{},{},{},{},{},{}\n'.format(folder,F1[0],F1[1],F1[2],F1[3],F1[4],F1[5]
,F1[6],F1[7],F1[8],F1[9],F1[10],F1[11],F1[12],F1[13],F1[14],F1[15],F1[16],F1[17]))
def compare_tnets_directed(th=50):
data_dir = 'dataset/'
folders = next(os.walk(data_dir))[1]
folders.sort()
TP_FP_FN_file = open(
'directed.tnet.old.new.th_' + str(th) + '.TP_FP_FN.csv', 'w+')
TP_FP_FN_file.write(
'dataset,tnet_old_tp,tnet_old_fp,tnet_old_fn,tnet_new_tp,tnet_new_fp,tnet_new_fn\n'
)
F1_file = open('directed.tnet.old.new.th_' + str(th) + '.F1.csv', 'w+')
F1_file.write(
'dataset,tnet_old_prec,tnet_old_rec,tnet_old_f1,tnet_new_prec,tnet_new_rec,tnet_new_f1\n'
)
for folder in folders:
print('inside folder: ', folder)
TP_FP_FN = []
F1 = []
real = set(
ge.get_real_edges(data_dir + folder + '/transmission_network.txt'))
tnet_old = set(
ge.get_mul_tnet_edges(data_dir + folder + '/tnet_old_100.tnet',
50))
tnet_new = set(
ge.get_mul_tnet_edges(data_dir + folder + '/tnet_new_100.tnet',
50))
TP = len(real & tnet_old)
FP = len(tnet_old - real)
FN = len(real - tnet_old)
try:
precision = TP / (TP + FP)
recall = TP / (TP + FN)
f1 = 2 * (recall * precision) / (recall + precision)
except ZeroDivisionError:
precision = 0
recall = 0
f1 = 0
TP_FP_FN.append(TP)
TP_FP_FN.append(FP)
TP_FP_FN.append(FN)
F1.append(round(precision, 3))
F1.append(round(recall, 3))
F1.append(round(f1, 3))
TP = len(real & tnet_new)
FP = len(tnet_new - real)
FN = len(real - tnet_new)
try:
precision = TP / (TP + FP)
recall = TP / (TP + FN)
f1 = 2 * (recall * precision) / (recall + precision)
except ZeroDivisionError:
precision = 0
recall = 0
f1 = 0
TP_FP_FN.append(TP)
TP_FP_FN.append(FP)
TP_FP_FN.append(FN)
F1.append(round(precision, 3))
F1.append(round(recall, 3))
F1.append(round(f1, 3))
TP_FP_FN_file.write('{},{},{},{},{},{},{}\n'.format(
folder, TP_FP_FN[0], TP_FP_FN[1], TP_FP_FN[2], TP_FP_FN[3],
TP_FP_FN[4], TP_FP_FN[5]))
F1_file.write('{},{},{},{},{},{},{}\n'.format(folder, F1[0], F1[1],
F1[2], F1[3], F1[4],
F1[5]))