def compute(compared_file, input_file):
save_result = 'output//Result_' + compared_file.split('\\')[-1].split(
'.')[0] + '_' + input_file.split('\\')[-1].split('.')[0] + '.xlsx'
try:
file = xlwt.Workbook()
sheet1 = file.add_sheet('similarity', cell_overwrite_ok=True)
except IOError:
return "Error: Failed to create file"
specialAlleleList = ['DYS385', 'DYF387S1']
sampleName_input, _, alleleList_input, data_input = readData.readExcelData(
input_file, specialAlleleList, SHUFFLE=False)
sampleName_compared, _, alleleList_compared, data_compared = readData.readExcelData(
compared_file, specialAlleleList, SHUFFLE=False)
# sheet1.write(0, 0, "")
for i in range(len(data_compared)):
sheet1.write(i + 1, 0, sampleName_compared[i])
for j in range(len(data_input)):
sheet1.write(0, j + 1, sampleName_input[j])
sheet1.write(
i + 1, j + 1,
str(round(cos_sim(data_compared[i], data_input[j]), 5)))
try:
file.save(save_result)
except IOError:
return "Error: File not found or failed to read "
else:
print(save_result + " has been written.")
return save_result + " has been written."
def sl_getDataFile(self):
text = self.ui.te_specialAllele.toPlainText()
global alleleList, file, exampleList
alleleList = []
file = []
exampleList = []
file, _ = QFileDialog.getOpenFileName(
self, "open example file", "",
"Excel File (*.xlsx);;Excel File (*.xls);;Text File(*.txt)")
if len(file) == 0:
return
self.ui.le_data.setText(file)
#check specialAllele isempty
if len(text) > 0:
specialAlleleList = re.split('[ ,.\n\t]', text)
else:
specialAlleleList = []
# specialAlleleList = []
#check type of file
typeFile = os.path.splitext(file)[-1]
if typeFile in ['.xlsx', '.xls']:
exampleList, alleleList, alleleData = readData.readExcelData(
file, specialAlleleList)
else:
exampleList, alleleList, alleleData = readData.readTxtData(
file, specialAlleleList)
self.grouping(exampleList, alleleData)
def compute(train_path, compared_file, input_file):
ml_method = [
'knn', 'naiveBayes', 'logisticRegression', 'svm', 'decesionTree',
'randomForest'
]
save_result = 'output//SimilarityResult_DATABASE_' + compared_file.split(
'\\')[-1].split('.')[0] + '_' + input_file.split('\\')[-1].split(
'.')[0] + '.xlsx'
try:
file = openpyxl.Workbook()
sheet1 = file.create_sheet('similarity')
except IOError:
return "Error: Failed to create file"
sampleName_input, _, alleleList_input, data_input = readData.readExcelData(
input_file, SHUFFLE=False)
sampleName_compared, _, alleleList_compared, data_compared = readData.readExcelData(
compared_file, SHUFFLE=False)
proba_compared = list()
proba_input = list()
for k in range(6):
model = os.path.join(train_path, ml_method[k] + '.model')
proba_input.append(load_compute(model, data_input))
proba_compared.append(load_compute(model, data_compared))
line_num = 1
sheet1.cell(line_num, 1).value = 'method'
for k in range(6):
sheet1.cell(line_num + 1, 1).value = ml_method[k]
for i in range(len(data_compared)):
sheet1.cell(line_num + 1, 2).value = sampleName_compared[i]
line_num += 1
for j in range(len(data_input)):
sheet1.cell(1, j + 3).value = sampleName_input[j]
sheet1.cell(line_num, j + 3).value = str(
round(cos_sim(proba_compared[k][i], proba_input[k][j]), 5))
try:
file.save(save_result)
except IOError:
return "Error: File not found or failed to read "
else:
print(save_result + " has been written.")
return save_result + " has been written."
def train(train_path):
len_train_file = len(os.listdir(train_path))
dirs = os.listdir(train_path)
specialAlleleList = ['DYS385', 'DYF387S1']
for file_num in range(len_train_file):
train_file = os.path.join(train_path, dirs[file_num])
save_model_path = 'output/train_model/' + train_path.split(
'\\')[-1] + '/'
if not os.path.exists(save_model_path):
os.makedirs(save_model_path)
alleleGroupName, alleleList, alleleData, data, groupName, shuffleIndex = readData.readExcelData(
train_file, specialAlleleList, SHUFFLE=True)
x_train = alleleData
y_train = alleleGroupName
#knn
knn = KNeighborsClassifier(1).fit(x_train, y_train)
joblib.dump(
knn, save_model_path + dirs[file_num].split('.')[0] + '_knn.model')
#naiveBayes
gnb = GaussianNB().fit(x_train, y_train)
joblib.dump(
gnb, save_model_path + dirs[file_num].split('.')[0] +
'_naiveBayes.model')
#logisticRegression
lr = LogisticRegression(penalty="l2",
C=1,
multi_class="ovr",
solver="newton-cg",
max_iter=1000).fit(x_train, y_train)
joblib.dump(
lr, save_model_path + dirs[file_num].split('.')[0] +
'_logisticRegression.model')
#SVM
svm_svm = svm.SVC(probability=True).fit(x_train, y_train)
joblib.dump(
svm_svm,
save_model_path + dirs[file_num].split('.')[0] + '_svm.model')
#decesionTree
dt = DecisionTreeClassifier(criterion='entropy').fit(x_train, y_train)
joblib.dump(
dt, save_model_path + dirs[file_num].split('.')[0] +
'_decesionTree.model')
#randomForest
rf = RandomForestClassifier(n_estimators=170).fit(x_train, y_train)
joblib.dump(
rf, save_model_path + dirs[file_num].split('.')[0] +
'_randomForest.model')
def test(mode, train_path, test_file):
ml_method = [
'knn', 'naiveBayes', 'logisticRegression', 'svm', 'decesionTree',
'randomForest'
]
len_train_file = len(os.listdir(train_path)) if mode == "others" else 1
dirs = os.listdir(train_path)
if (mode == "DATABASE"):
save_model_path = train_path + '/'
save_result = 'output/PredictResult_DATABASE_' + test_file.split(
'\\')[-1].split('.')[0] + '.xlsx'
else:
save_model_path = 'output/train_model/' + train_path.split(
'\\')[-1] + '/'
if not os.path.exists(save_model_path):
os.makedirs(save_model_path)
save_result = 'output/PredictResult_' + train_path.split(
'\\')[-1] + '_' + test_file.split('\\')[-1].split('.')[0] + '.xlsx'
try:
file = openpyxl.Workbook()
sheet1 = file.create_sheet('predict')
except IOError:
return "Error: Failed to create file"
sampleName_test, alleleGroupName_test, alleleList_test, data_test = readData.readExcelData(
test_file, SHUFFLE=False)
line_num = 0
for file_num in range(len_train_file):
sheet1.cell(
line_num + 1,
0 + 1).value = dirs[file_num] if mode == "others" else "DATABASE"
line_num += 1
for data_num in range(len(data_test)):
sheet1.cell(line_num + 1, data_num + 1 + 1).value = str(
sampleName_test[data_num])
line_num += 1
x_test = data_test
for k in range(6):
if (mode == "others"):
model = os.path.join(
save_model_path, dirs[file_num].split('.')[0] + '_' +
ml_method[k] + '.model')
else:
model = os.path.join(save_model_path, ml_method[k] + '.model')
try:
fout = open(model)
except:
print(model + ' do not exist. Please train first.')
return
sheet1, line_num = load_pred(model, x_test, len(data_test), sheet1,
line_num, ml_method[k])
# #knn
# model = os.path.join(save_model_path,dirs[file_num].split('.')[0]+'_knn.model')
#
# #naiveBayes
# model = os.path.join(save_model_path,dirs[file_num].split('.')[0]+'_naiveBayes.model')
# sheet1, line_num=load_pred(mode,model,train_path,x_test,len(data_test),sheet1,line_num,"naiveBayes")
#
# #logisticRegression
# model = os.path.join(save_model_path, dirs[file_num].split('.')[0] + '_logisticRegression.model')
# sheet1, line_num=load_pred(mode,model,train_path,x_test,len(data_test),sheet1,line_num,"logisticRegression")
#
# #SVM
# model = os.path.join(save_model_path, dirs[file_num].split('.')[0] + '_svm.model')
# sheet1, line_num=load_pred(mode,model,train_path,x_test,len(data_test),sheet1,line_num,"svm")
#
# #decesionTree
# model = os.path.join(save_model_path, dirs[file_num].split('.')[0] + '_decesionTree.model')
# sheet1, line_num=load_pred(mode,model,train_path,x_test,len(data_test),sheet1,line_num,"decesionTree")
#
# #randomForest
# model = os.path.join(save_model_path, dirs[file_num].split('.')[0] + '_randomForest.model')
# sheet1, line_num=load_pred(mode,model,train_path,x_test,len(data_test),sheet1,line_num,"randomForest")
try:
file.save(save_result)
except IOError:
return "Error: File not found or failed to read "
else:
print(save_result + " has been written.")
return save_result + " has been written."
def func(train_path, test_file):
#readdata
len_train_file = len(os.listdir(train_path))
dirs = os.listdir(train_path)
# write
try:
file = xlwt.Workbook()
sheet1 = file.add_sheet('predict', cell_overwrite_ok=True)
sheet2 = file.add_sheet('similarity', cell_overwrite_ok=True)
except IOError:
return "Error: Failed to create file"
score_result = [[0 for i in range(6)] for j in range(len_train_file)]
specialAlleleList = ['DYS385', 'DYF387S1']
alleleGroupName_test, alleleList_test, data_test = readData.readExcelData(
test_file, specialAlleleList, SHUFFLE=False)
result = [[] for _ in range(len_train_file)]
line_num = 0
for file_num in range(1):
sheet1.write(line_num, 0, dirs[file_num])
sheet2.write(line_num, 0, dirs[file_num])
line_num += 1
train_file = os.path.join(train_path, dirs[file_num])
print(train_file)
alleleGroupName, alleleList, alleleData, data, groupName, shuffleIndex = readData.readExcelData(
train_file, specialAlleleList, SHUFFLE=True)
compare_result = [[] for i in range(6)]
original_compare_result = [[] for i in range(len(data))]
loop_num = 1
x_train = alleleData
x_test = data_test
y_train = alleleGroupName
unique_group = np.unique(y_train)
#knn
knn = KNeighborsClassifier(1).fit(x_train, y_train)
knn_pred = knn.predict(x_test)
knn_proba = knn.predict_proba(x_test)
result[file_num].append(knn_pred)
for jj in range(len(data_test)):
# sheet1.write(line_num, jj, knn_pred[jj])
sheet1.write(
line_num, jj,
str(knn_pred[jj]) + '(' + str(round(max(knn_proba[jj]), 3)) +
')')
for jj in range(1, len(data_test)):
sheet2.write(line_num, jj - 1, cos_sim(knn_proba[jj],
knn_proba[0]))
line_num += 1
#naiveBayes
gnb = GaussianNB().fit(x_train, y_train)
# predict_score = gnb.predict_proba(x_test)
gbn_pred = gnb.predict(x_test)
gbn_proba = gnb.predict_proba(x_test)
result[file_num].append(gbn_pred)
for jj in range(len(data_test)):
sheet1.write(
line_num, jj,
str(gbn_pred[jj]) + '(' + str(round(max(gbn_proba[jj]), 3)) +
')')
for jj in range(1, len(data_test)):
sheet2.write(line_num, jj - 1, cos_sim(gbn_proba[jj],
gbn_proba[0]))
line_num += 1
#logisticRegression
lr = LogisticRegression(penalty="l2",
C=1,
multi_class="ovr",
solver="newton-cg",
max_iter=1000).fit(x_train, y_train)
# print(lr.predict_proba(x_test))
lr_pred = lr.predict(x_test)
lr_proba = lr.predict_proba(x_test)
result[file_num].append(lr_pred)
for jj in range(len(data_test)):
sheet1.write(
line_num, jj,
str(lr_pred[jj]) + '(' + str(round(max(lr_proba[jj]), 3)) +
')')
for jj in range(1, len(data_test)):
sheet2.write(line_num, jj - 1, cos_sim(lr_proba[jj], lr_proba[0]))
line_num += 1
#SVM
svm_svm = svm.SVC(probability=True)
svm_svm.fit(x_train, y_train)
svm_pred = svm_svm.predict(x_test)
svm_proba = svm_svm.predict_proba(x_test)
result[file_num].append(svm_pred)
for jj in range(len(data_test)):
sheet1.write(
line_num, jj,
str(svm_pred[jj]) + '(' + str(round(max(svm_proba[jj]), 3)) +
')')
for jj in range(1, len(data_test)):
sheet2.write(line_num, jj - 1, cos_sim(svm_proba[jj],
svm_proba[0]))
line_num += 1
#decesionTree
dt = DecisionTreeClassifier(criterion='entropy')
dt.fit(x_train, y_train)
dt_pred = dt.predict(x_test)
dt_proba = dt.predict_proba(x_test)
result[file_num].append(dt_pred)
for jj in range(len(data_test)):
sheet1.write(
line_num, jj,
str(svm_pred[jj]) + '(' + str(round(max(dt_proba[jj]), 3)) +
')')
# sheet1.write(line_num, jj, svm_pred[jj])
for jj in range(1, len(data_test)):
sheet2.write(line_num, jj - 1, cos_sim(dt_proba[jj], dt_proba[0]))
line_num += 1
#randomForest
rf = RandomForestClassifier(n_estimators=170)
rf.fit(x_train, y_train)
rf_pred = rf.predict(x_test)
rf_proba = rf.predict_proba(x_test)
result[file_num].append(rf_pred)
for jj in range(len(data_test)):
sheet1.write(
line_num, jj,
str(rf_pred[jj]) + '(' + str(round(max(rf_proba[jj]), 3)) +
')')
for jj in range(1, len(data_test)):
sheet2.write(line_num, jj - 1, cos_sim(rf_proba[jj], rf_proba[0]))
line_num += 1
loop_num += 1
try:
file.save('output//Result_' + train_path.split('\\')[-1] + '_' +
test_file.split('\\')[-1].split('.')[0] + '.xlsx')
except IOError:
return "Error: File not found or failed to read "
else:
print("'result.xlsx' has been written.")
return "'result.xlsx' has been written."
def test(train_path, test_file):
len_train_file = len(os.listdir(train_path))
dirs = os.listdir(train_path)
save_model_path = 'output/train_model/' + train_path.split('\\')[-1] + '/'
save_result = 'output//Result_' + train_path.split(
'\\')[-1] + '_' + test_file.split('\\')[-1].split('.')[0] + '.xlsx'
try:
file = xlwt.Workbook()
sheet1 = file.add_sheet('predict', cell_overwrite_ok=True)
# sheet2 = file.add_sheet('similarity', cell_overwrite_ok=True)
except IOError:
return "Error: Failed to create file"
specialAlleleList = ['DYS385', 'DYF387S1']
sampleName_test, alleleGroupName_test, alleleList_test, data_test = readData.readExcelData(
test_file, specialAlleleList, SHUFFLE=False)
line_num = 0
for file_num in range(len_train_file):
sheet1.write(line_num, 0, dirs[file_num])
line_num += 1
# sheet2.write(line_num, 0, dirs[file_num])
for data_num in range(len(data_test)):
sheet1.write(line_num, data_num + 1,
str(sampleName_test[data_num]))
# sheet2.write(line_num, data_num+1, str(sampleName_test[data_num]))
line_num += 1
x_test = data_test
#knn
model = os.path.join(save_model_path,
dirs[file_num].split('.')[0] + '_knn.model')
sheet1, line_num = load_pred(model, train_path, x_test, len(data_test),
sheet1, line_num, "knn")
#naiveBayes
model = os.path.join(
save_model_path,
dirs[file_num].split('.')[0] + '_naiveBayes.model')
sheet1, line_num = load_pred(model, train_path, x_test, len(data_test),
sheet1, line_num, "naiveBayes")
#logisticRegression
model = os.path.join(
save_model_path,
dirs[file_num].split('.')[0] + '_logisticRegression.model')
sheet1, line_num = load_pred(model, train_path, x_test, len(data_test),
sheet1, line_num, "logisticRegression")
#SVM
model = os.path.join(save_model_path,
dirs[file_num].split('.')[0] + '_svm.model')
sheet1, line_num = load_pred(model, train_path, x_test, len(data_test),
sheet1, line_num, "svm")
#decesionTree
model = os.path.join(
save_model_path,
dirs[file_num].split('.')[0] + '_decesionTree.model')
sheet1, line_num = load_pred(model, train_path, x_test, len(data_test),
sheet1, line_num, "decesionTree")
#randomForest
model = os.path.join(
save_model_path,
dirs[file_num].split('.')[0] + '_randomForest.model')
sheet1, line_num = load_pred(model, train_path, x_test, len(data_test),
sheet1, line_num, "randomForest")
try:
file.save(save_result)
except IOError:
return "Error: File not found or failed to read "
else:
print(save_result + " has been written.")
return save_result + " has been written."