def __init__(self, path):
self.input_file = path
self.preprocesser = preprocessar()
self.cluster = cluter()
self.best_model = best_model_tuner()
self.log = APP_logger()
self.model_save = model_file()
class model_file:
def __init__(self):
self.file = "model//"
self.log = APP_logger()
self.pickle = pickle
def model_save(self, model, filename):
try:
path = self.file + filename
if os.path.isdir(path):
shutil.rmtree(path)
os.makedirs(path)
else:
os.makedirs(path)
with open(path + "//" + filename + ".pickle", "wb") as f:
pickle.dump(model, f)
f.close()
message = "{a} model saved succesfully".format(a=model)
self.log.log("log_msg//model_load_pre.txt", message)
except Exception as e:
message = "there is some erroe in this funtion,the error is {fun}".format(
fun=e)
self.log.log("log_msg//model_load_pre.txt", message)
return e
def load_model(self, filename):
try:
self.file = filename
path = "model//" + self.file
with open(path + "//" + self.file + ".pickle", "rb") as f:
model = pickle.load(f)
f.close()
message = "successfully {model} imported ".format(model=filename)
self.log.log("log_msg//load_mdel.txt", message)
return model
except Exception as e:
message = "there is some erroe in this funtion,the error is {fun}".format(
fun=e)
self.log.log("log_msg//load_model.txt", message)
return e
def find_correct_model_file(self, num_clu):
#try:
list_file = os.listdir(self.file)
for file in list_file:
if file[-1] == str(num_clu):
return file
message = "successfully runed this programme"
self.log.log("log_msg//model_best_pre..txt", message)
class model_training:
def __init__(self, path):
self.input_file = path
self.preprocesser = preprocessar()
self.cluster = cluter()
self.best_model = best_model_tuner()
self.log = APP_logger()
self.model_save = model_file()
def training_model(self):
#try:
data = pd.read_csv(self.input_file)
X, Y = self.preprocesser.splite_data(data,
"default payment next month")
# x_train,x_test,y_train,y_test=self.preprocesser.train_test_splt(X,Y)
n_cluster = self.cluster.num_cluster(X)
x = self.cluster.K_cluster(X, n_cluster)
self.model_save.model_save(self.cluster, "KMeans")
x["label"] = Y
list_of_cluster = x["cluster"].unique()
for i in list_of_cluster:
cluster_data = X[X["cluster"] == i]
cluster_feature = cluster_data.drop(columns=["cluster", "label"])
cluster_label = cluster_data["label"]
x_train, x_test, y_train, y_test = train_test_split(
cluster_feature, cluster_label, test_size=.30, random_state=42)
model, model_name = self.best_model.best_model(
x_train, x_test, y_train, y_test)
self.model_save.model_save(model, model_name + str(i))
message = "succesfully trainied model"
self.log.log("log_msg//model_finaltrain.txt", message)
def __init__(self):
self.xgboost = XGBClassifier()
self.naive = GaussianNB()
self.log = APP_logger()
class best_model_tuner:
def __init__(self):
self.xgboost = XGBClassifier()
self.naive = GaussianNB()
self.log = APP_logger()
def best_para_for_naive(self, x_train, y_train):
try:
self.para_gired = {
"var_smoothing":
[1e-9, .001, .05, .08, .1, .5, 1e-8, 1e-7, 1e-6]
}
giried = GridSearchCV(self.naive, self.para_gired, cv=5, verbose=3)
giried.fit(x_train, y_train)
best_var_smoothing = giried.best_params_["var_smoothing"]
# return best_var_smoothing
self.naive1 = GaussianNB(var_smoothing=best_var_smoothing)
self.naive1.fit(x_train, y_train)
message = "succesfully naive model best parameter find"
self.log.log("log_msg//naive_model.txt", message)
return self.naive1
except Exception as e:
message = "there is some error in database creation function ,the error is {a}".format(
a=e)
self.log.log("log_msg//naive_model.txt", message)
return e
def best_para_for_XGboost(self, x_train, y_train):
try:
para = {"n_estimators": [100, 200, 250], "max_depth": [2, 6, 10]}
gired = GridSearchCV(self.xgboost,
param_grid=para,
cv=5,
verbose=3)
gired.fit(x_train, y_train)
# return gired.best_params_
self.n_estimater = gired.best_params_["n_estimators"]
self.max_depth = gired.best_params_["max_depth"]
self.xgboost = XGBClassifier(n_estimators=self.n_estimater,
max_depth=self.max_depth)
self.xgboost.fit(x_train, y_train)
message = "succesfully xgboost model best parameter find"
self.log.log("log_msg//xgboost_model.txt", message)
return self.xgboost
except Exception as e:
message = "there is some error in database creation function ,the error is {a}".format(
a=e)
self.log.log("log_msg//xgboost_model.txt", message)
return e
def best_model(self, x_train, x_test, y_train, y_test):
try:
self.xgboost = self.best_para_for_XGboost(x_train, y_train)
self.naive = self.best_para_for_naive(x_train, y_train)
xg_pre = self.xgboost.predict(x_test)
naive_pre = self.naive.predict(x_test)
xg_score = accuracy_score(y_test, xg_pre)
naive_score = accuracy_score(y_test, naive_pre)
message = "succesfully best model find"
self.log.log("log_msg//best_model.txt", message)
# return naive_score,xg_score
if xg_score > naive_score:
return self.xgboost, "xgboost"
else:
return self.naive, "naive"
except Exception as e:
message = "there is some error in database creation function ,the error is {a}".format(
a=e)
self.log.log("log_msg//best_model.txt", message)
return e
class db_vali:
def __init__(self):
self.path = "Training_database//"
self.log = APP_logger()
self.good_file_path = "Training_raw_file//Good_data"
self.bad_file_path = "Training_raw_file//bad_data"
def DBconnection(self, DatabaseName):
try:
conn = sqlite3.connect(self.path + DatabaseName + ".db")
message = "succesfully Database connection done"
self.log.log("log_msg//database.txt", message)
except Exception as e:
message = "there is some error in database creation function ,the error is {a}".format(a=e)
self.log.log("log_msg//database.txt", message)
return e
return conn
def dbtablecreation(self, Databasename, column):
try:
conn = self.DBconnection(Databasename)
c = conn.cursor()
c.execute("SELECT count(name) FROM sqlite_master WHERE type = 'table'AND name = 'Good_Raw_Data'")
if c.fetchone()[0] == 1:
conn.close()
else:
for col in column.keys():
type=column[col]
try:
c.execute("ALTER TABLE Good_Raw_Data ADD column'{col_name}'{type}".format(col_name=col,type=type))
except:
c.execute("CREATE TABLE Good_Raw_Data({col_name} {type})".format(col_name=col,type=type))
message = "succesfully Database table_creation connection done"
self.log.log("log_msg//database.txt", message)
except Exception as e:
message = "there is some error in database creation function ,the error is {a}".format(a=e)
self.log.log("log_msg//database.txt", message)
return e
conn.close()
def insert_table_into_db(self,Databasename):
try:
conn=self.DBconnection(Databasename)
c=conn.cursor()
list_files=os.listdir("training_raw_file//Good_data//")
for i in list_files:
with open("training_raw_file//Good_data//"+i,"r")as f:
next(f,1)
for j in enumerate(f):
c.execute("INSERT INTO Good_Raw_Data VALUES ({VALUE})".format(VALUE=j[1]))
message = "succesfully Database connection done"
self.log.log("log_msg//database.txt", message)
conn.commit()
c.close()
conn.close()
except Exception as e:
message = "there is some error in database creation function ,the error is {a}".format(a=e)
self.log.log("log_msg//database.txt", message)
return e
conn.close()
def Export_DB_csv(self,Databasename):
try:
input_file = "FileFrom_DB"
conn = self.DBconnection(Databasename)
c = conn.cursor()
if not os.path.isdir(input_file):
os.makedirs(input_file)
c.execute("SELECT * FROM Good_Raw_Data")
result=c.fetchall()
header=[i[0] for i in c.description]
csv_file=csv.writer(open("FileFrom_DB//input.csv","w",newline=""),delimiter=",")
csv_file.writerow(header)
csv_file.writerows(result)
message = "succesfully Database Exported into csv file connection done"
self.log.log("log_msg//database.txt", message)
except Exception as e:
message = "there is some error in database creation function ,the error is {a}".format(a=e)
self.log.log("log_msg//database.txt", message)
c.close()
def __init__(self):
self.path = "Training_database//"
self.log = APP_logger()
self.good_file_path = "Training_raw_file//Good_data"
self.bad_file_path = "Training_raw_file//bad_data"
def __init__(self):
self.path = "prediction_database//"
self.log = APP_logger()
self.good_file_path = "prediction_raw_file//Good_data"
self.bad_file_path = "prediction_raw_file//bad_data"
class prediction_val:
def __init__(self,prediction_file_dir):
self.Batch_dir=prediction_file_dir
self.Log=APP_logger()
self.schema="schema_prediction.json"
def schema_values(self):
try:
with open(self.schema,"r")as f:
dic=json.load(f)
SampleFileName=dic["SampleFileName"]
LengthOfDateStampInFile=dic["LengthOfDateStampInFile"]
LengthOfTimeStampInFile=dic["LengthOfTimeStampInFile"]
NumberofColumns=dic["NumberofColumns"]
ColName=dic["ColName"]
message="succesfully schema_value created"
self.Log.log("log_msg//schema_values_pre.txt",message)
return SampleFileName,LengthOfDateStampInFile,LengthOfTimeStampInFile,NumberofColumns,ColName
except Exception as e:
message = "there is some error schema_values function ,the error is {a}".format(a=e)
self.Log.log("log_msg//schema_values_pre.txt", message)
return e
def create_good_bad_folder(self):
try:
path=os.path.join("prediction_raw_file//Good_data")
if not os.path.isdir(path):
os.makedirs(path)
path=os.path.join("prediction_raw_file//bad_data")
if not os.path.isdir(path):
os.makedirs(path)
message = "succesfully created good and bad raw data folder"
self.Log.log("log_msg//create_good_bad_pre.txt", message)
except Exception as e:
message = "there is some error in create good and bad data folder function ,the error is {a}".format(a=e)
self.Log.log("log_msg//create_good_bad_pre.txt", message)
return e
def delete_good_data(self):
try:
path="prediction_raw_file//Good_data"
if os.path.isdir(path):
shutil.rmtree(path)
message = "succesfully deleted good raw data folder"
self.Log.log("log_msg//delete_good_pre.txt", message)
except Exception as e:
message = "there is some error in delete good data folder function ,the error is {a}".format(a=e)
self.Log.log("log_msg//delete_good_pre.txt", message)
return e
def delete_bad_data(self):
try:
path="prediction_raw_file//bad_data"
if os.path.isdir(path):
shutil.rmtree(path)
message = "succesfully deleted bad raw data folder"
self.Log.log("log_msg//delete_bad_pre.txt", message)
except Exception as e:
message = "there is some error in delete bad data folder function ,the error is {a}".format(a=e)
self.Log.log("log_msg//delete_bad_pre.txt", message)
return e
def move_bad_to_archivedbad(self):
try:
now=datetime.now()
date=now.date()
current_time=date.strftime("%H%M%S")
source="prediction_raw_file//bad_data"
if os.path.isdir(source):
path="prediction_Archived"
if not os.path.isdir(path):
os.makedirs(path)
source1=path+"\\bad_data_"+str(date)+"-"+str(current_time)
if not os.path.isdir(source1):
os.makedirs(source1)
list_files=os.listdir(source)
for file in list_files:
if file not in os.listdir(source1):
shutil.move(source+"//"+file,source1)
message = "succesfully moved bad_file to Archivedbad folder raw data folder"
self.Log.log("log_msg//Archived_bad_pre.txt", message)
except Exception as e:
message = "there is some error in move_bad_to_archivedbad function ,the error is {a}".format(a=e)
self.Log.log("log_msg//Archived_bad_pre.txt", message)
return e
def cheack_filename(self,len_date,lentime):
try:
list_files=os.listdir("prediction_Batch_Files")
for file in list_files:
f_split=re.split(".csv",file)[0]
s_split=re.split("_",f_split)
if (s_split[0]=="creditCardFraud") and (len(s_split[1])==len_date) and (len(s_split[2])==lentime):
shutil.copy("prediction_Batch_Files//"+file,"prediction_raw_file//Good_data")
else:
shutil.copy("prediction_Batch_Files//" + file, "prediction_raw_file//bad_data")
message = "succesfully checked all the correct file and copyed to respect folders"
self.Log.log("log_msg//check_filename_.txt", message)
except Exception as e:
message = "there is having some error in file checked function ,the error is {a}".format(a=e)
self.Log.log("log_msg//check_filename.txt", message)
return e
def cheack_col(self,numofcol):
try:
list_files=os.listdir("prediction_raw_file//Good_data")
for file in list_files:
data=pd.read_csv("prediction_raw_file//Good_data//"+file)
if data.shape[1]==numofcol:
pass
else:
shutil.move("prediction_raw_file//Good_data//"+file,"prediction_raw_file//bad_data")
message = "succesfully checked all columns and copyed to the respect folders"
self.Log.log("log_msg//check_column_pre.txt", message)
except Eception as e:
message = "there is having some error in column checked function ,the error is {a}".format(a=e)
self.Log.log("log_msg//check_column_pre.txt", message)
return e
def cheack_mis_col(self):
try:
list_files = os.listdir("prediction_raw_file//Good_data")
for file in list_files:
data=pd.read_csv("prediction_raw_file//Good_data//"+file)
for col in data.columns:
if len(data[col])-data[col].count()==len(data[col]):
shutil.move("prediction_raw_file//Good_data//"+file,"prediction_raw_file//bad_data")
message = "succesfully checked all columns and copyed to the respect folders"
self.Log.log("log_msg//check_missing_column_pre.txt", message)
except Exception as e:
message = "there is having some error in this function ,the error is {a}".format(a=e)
self.Log.log("log_msg//check_missing_column_pre.txt", message)
return e
def __init__(self,prediction_file_dir):
self.Batch_dir=prediction_file_dir
self.Log=APP_logger()
self.schema="schema_prediction.json"
def __init__(self, path):
self.path = path
self.prediction_val = prediction_val(self.path)
self.db_vali = db_vali()
self.log = APP_logger()
def __init__(self, path):
self.path = path
self.Training_val = Training_val(self.path)
self.db_vali = db_vali()
self.log = APP_logger()
def __init__(self):
self.file = "model//"
self.log = APP_logger()
self.pickle = pickle
def __init__(self, training_file_dir):
self.Batch_dir = training_file_dir
self.Log = APP_logger()
self.schema = "schema_training.json"
