def main():
print("have a test")
print(os.path)
print(os.getcwd())
data_in = pd.read_csv(os.path.join(data_map.datain, 'german_credit_data.csv'))
print(data_in.info())
print(data_in.head())
clean_data = data_process.data_clean(data_in)
print(clean_data.info())
print(clean_data.head())
#clean_data.to_csv(os.path.join(data_map.dataout, 'clean_data.csv'), index=False,encoding='utf-8')
"""
X, y = data_process.data_trans(clean_data)
print(X.shape)
print(y.shape)
"""
train_data, test_data = train_test_split(clean_data, test_size=1/4, random_state=10)
data_process.dataset_view(train_data,test_data)
X_train, y_train = data_process.data_trans(train_data)
X_test, y_test = data_process.data_trans(test_data)
model_name_para_dic = {'kNN':[5,10,15],
'LR':[0.01,1,100]}
results_df = pd.DataFrame(columns=['Accuracy (%)', 'Time (s)'],
index=list(model_name_para_dic.keys()))
results_df.index.name = 'Model'
for model_name, para in model_name_para_dic.items():
#print(model_name,para)
best_acc, best_model, mean_duration = data_process.train_test_model(X_train,y_train,X_test,y_test,para,model_name)
#print(best_acc*100)
#print(best_model)
#print(mean_duration)
results_df.loc[model_name, 'Accuracy (%)'] = best_acc * 100
results_df.loc[model_name, 'Time (s)'] = mean_duration
plt.figure()
ax1 = plt.subplot(1, 2, 1)
results_df.plot(y=['Accuracy (%)'], kind='bar', ylim=[60, 100], ax=ax1, title='Accuracy(%)', legend=False)
ax2 = plt.subplot(1, 2, 2)
results_df.plot(y=['Time (s)'], kind='bar', ax=ax2, title='Time(s)', legend=False)
plt.tight_layout()
plt.savefig(os.path.join(data_map.dataout, 'result.png'))
from sklearn.cross_validation import train_test_split
from sklearn import metrics
import tensorflow as tf
from data_process import data_clean
full_X, full = data_clean()
sourceRow = 891 # 训练数据集的大小
# 原始数据集：特征
source_X = full_X.loc[0:sourceRow - 1, :]
# 原始数据集：标签
source_y = full.loc[0:sourceRow - 1, 'Survived']
# 预测数据集：特征
pred_X = full_X.loc[sourceRow:, :]
train_X, test_X, train_Y, test_Y = train_test_split(source_X,
source_y,
train_size=.75)
feature_columns = [tf.contrib.layers.real_valued_column("", dimension=27)]
model = tf.contrib.learn.DNNClassifier(hidden_units=[10, 20, 10],
feature_columns=feature_columns)
model.fit(test_X, test_Y)
score = metrics.accuracy_score(test_Y, model.predict(test_X))
print(score)
def run_it():
#start time (IST)
start_time = (datetime.datetime.utcnow() +
datetime.timedelta(hours=5, minutes=30)).isoformat()
#instruments_list
nse_instruments, bse_instruments = data_process.instruments_list()
#sockets and quote_types variables
nse_sockets = []
bse_sockets = []
nse_quote_type = [
'nse-m',
'nse-m',
'nse-m',
'nse-m',
'nse-m',
'nse-m',
'nse-s',
'nse-s',
'nse-s',
'nse-s',
'nse-s',
'nse-s',
]
bse_quote_type = [
'bse-m',
'bse-m',
'bse-m',
'bse-m',
'bse-m',
'bse-m',
'bse-s',
'bse-s',
'bse-s',
'bse-s',
'bse-s',
'bse-s',
]
#socket object creation
for i in range(0, 12):
nse_sockets.append(alice_obj.socket(nse_quote_type[i]))
bse_sockets.append(alice_obj.socket(bse_quote_type[i]))
#subscription
for i in range(0, 12):
j = i
i = i % 6 #instruments are shared between market and snapquote sockets
if (i + 1) * 250 > len(nse_instruments) or (i + 1) * 250 > len(
bse_instruments):
nse_sockets[j].subscribe(nse_instruments[250 *
i:len(nse_instruments)])
bse_sockets[j].subscribe(bse_instruments[250 *
i:len(bse_instruments)])
else:
nse_sockets[j].subscribe(nse_instruments[250 * i:250 * (i + 1)])
bse_sockets[j].subscribe(bse_instruments[250 * i:250 * (i + 1)])
#sleep for data collection; As of now it'll be 10 mins
time.sleep(5)
#data_clean ( exchange variable for file_name)
nse_file_name = data_process.data_clean(nse_sockets, 'nse')
bse_file_name = data_process.data_clean(bse_sockets, 'bse')
#upload
upload.upload(nse_file_name)
upload.upload(bse_file_name)
#delete those files after upload
os.remove(nse_file_name)
os.remove(bse_file_name)
#endtime (IST)
end_time = (datetime.datetime.utcnow() +
datetime.timedelta(hours=5, minutes=30)).isoformat()
#work_logs to be returned to flask's @app.route('/')
work_logs = ' (IST) ' + 'start : ' + start_time + ' end : ' + end_time
return work_logs