def test_basic_experiment(): pa=os.getcwd() config_name='UCLSE\\test\\fixtures\\mkt_cfg.yml' config_path=os.path.join(pa,config_name) environ_dic=yamlLoad(config_path) #environ_dic['trade_file']=os.path.join(pa,'UCLSE\\test\\output\\avg_balance.csv') sess=Market_session(**environ_dic) sess.simulate(sess.trade_stats_df3)
def test_forwards_multi_leg():
pa=os.getcwd()
config_name='UCLSE\\test\\fixtures\\mkt_cfg.yml'
config_path=os.path.join(pa,config_name)
environ_dic=yamlLoad(config_path)
environ_dic['end_time']=50
sess=Market_session(**environ_dic)
sess1=copy.deepcopy(sess)
#sess1.process_order=sess1.exchange.process_order3w
try:
side_by_side_period_by_period_difference_checker(sess,sess1)
except AssertionError:
print('sess1 uses new process order,sess old')
raise
def test_replay_for_same_results():
pa=os.getcwd()
config_name='UCLSE\\test\\fixtures\\mkt_cfg.yml'
config_path=os.path.join(pa,config_name)
environ_dic=yamlLoad(config_path)
environ_dic['end_time']=50
sess=Market_session(**environ_dic)
sess1=copy.deepcopy(sess)
sess.simulate(sess.trade_stats_df3,recording=True)
reset_trader_class_variables(sess1)
sess1.simulate(sess1.trade_stats_df3,replay_vars=sess.replay_vars,recording=True)
assert identical_replay_vars(sess,sess1)=={}
assert (sess1.df.fillna(0)==sess.df.fillna(0)).all().all()
def build_df_from_yaml(path):
## builds a df from a yaml file containing any number of sub dictionaries
#of the form {'tid':[],'otype':[],'price':[],'qty':[],'time':[],'qid':[]}
#typically but not necessarily one dictionary for bids, one for asks.
necessary_cols = ['tid', 'otype', 'price', 'qty', 'time', 'qid']
dic = yamlLoad(path)
try:
for _, k in dic.items():
for col in necessary_cols:
assert col in k
except AssertionError:
print('All of ', necessary_cols, ' must be in each sub dictionary')
order_df = build_df_from_dic_dic(dic)
order_df = order_df[necessary_cols]
return order_df
def test_multi_q_exchange():
pa=os.getcwd()
config_name='UCLSE\\test\\fixtures\\mkt_cfg.yml'
config_path=os.path.join(pa,config_name)
environ_dic=yamlLoad(config_path)
environ_dic['end_time']=50
environ_dic['trader_record']=True
def geometric_q():
return np.random.geometric(0.6)
environ_dic['quantity_f']=geometric_q
sess=Market_session(**environ_dic)
sess1=copy.deepcopy(sess)
try:
side_by_side_period_by_period_difference_checker(sess,sess1)
except AssertionError:
print('multi q exchange failure')
raise
_test_trader_attributes(sess)
def test_multi_q_exchange_order_quantity():
#this one checks that no trades are executed more than the original order specified
pa=os.getcwd()
config_name='UCLSE\\test\\fixtures\\mkt_cfg.yml'
config_path=os.path.join(pa,config_name)
environ_dic=yamlLoad(config_path)
environ_dic['end_time']=50
def geometric_q():
return np.random.geometric(0.6)
environ_dic['quantity_f']=geometric_q
sess=Market_session(**environ_dic)
sess.simulate()
listy=[]
for t in sess.traders:
listy=listy+sess.traders[t].blotter
#listy[0]
df=pd.DataFrame(listy).groupby(['tid','oid']).agg({'order qty':'max','qty':'sum'})
assert (df.qty<=df['order qty']).all()
# def test_trade_stats_methods():
# #checks to see if the df is the same as the old fashioned csv print method
# pa=os.getcwd()
# config_name='UCLSE\\test\\fixtures\\mkt_cfg.yml'
# config_path=os.path.join(pa,config_name)
# environ_dic=yamlLoad(config_path)
# environ_dic['end_time']=50
# sess=Market_session(**environ_dic)
# sess1=copy.deepcopy(sess)
# sess.simulate(sess.trade_stats_df3,recording=True)
# reset_trader_class_variables(sess1)
# sess1.simulate(sess1.trade_stats,replay_vars=sess.replay_vars,recording=True)
# cols=pd.MultiIndex(levels=[['GVWY', 'SHVR', 'ZIC', 'ZIP', 'best_ask', 'best_bid', 'expid', 'time'], ['', 'balance_sum', 'n', 'pc']],
# codes=[[6, 7, 0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 5, 4], [0, 0, 1, 2, 3, 1, 2, 3, 1, 2, 3, 1, 2, 3, 0, 0]])
# d=pd.read_csv(environ_dic['trade_file'],header=None,usecols=[0,1,3,4,5,7,8,9,11,12,13,15,16,17,18,19])
# d.columns=cols
# #the csv method truncates accuracy of time value
# d.set_index(sess.df.index,inplace=True)
# cols=pd.MultiIndex(levels=[['GVWY', 'SHVR', 'ZIC', 'ZIP', 'best_ask', 'best_bid'], ['', 'balance_sum', 'n', 'pc']],
# codes=[[ 0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 5, 4], [ 1, 2, 3, 1, 2, 3, 1, 2, 3, 1, 2, 3, 0, 0]])
# #check the data is the same
# assert (d[cols].replace(' N',0).astype('float64')==sess.df[cols].fillna(0)).all().all()
# #check the time periods are the same
# assert np.allclose(d.index.values-sess.df.index.values,np.zeros(d.index.values.shape))