def initialize(context):
""" Called once at the start of the algorithm. """
set_slippage(slippage.VolumeShareSlippage(volume_limit=0.025, price_impact=0.1))
set_commission(commission.PerShare(cost=0.01, min_trade_cost=1.00))
set_max_leverage(1.0)
# Rebalance every day, 1 hour after market open.
schedule_function(
context.my_rebalance,
date_rules.every_day(),
time_rules.market_open(hours=1)
)
# Close all positions every day, 30 minutes before market close.
schedule_function(
context.close_positions,
date_rules.every_day(),
time_rules.market_close(minutes=30)
)
# Create risk manager
context.risk_manager = RiskManager(context, daily_risk)
# Create our dynamic stock selector.
attach_pipeline(context.make_screener(), 'stock_screener')
def initialize(context):
context.asset = symbol('AAPL')
# Explicitly set the commission/slippage to the "old" value until we can
# rebuild example data.
# github.com/quantopian/zipline/blob/master/tests/resources/
# rebuild_example_data#L105
context.set_commission(commission.PerShare(cost=.0075, min_trade_cost=1.0))
context.set_slippage(slippage.VolumeShareSlippage())
def initialize(context):
context.asset = symbol(asset_symbol)
# To keep track of whether we invested in the stock or not
context.invested = False
# Explicitly set the commission/slippage to the "old" value until we can
context.set_commission(commission.PerShare(**commission_cost))
context.set_slippage(slippage.VolumeShareSlippage())
context.params = params_list
def initialize(context):
"""Setup: register pipeline, schedule rebalancing,
and set trading params"""
attach_pipeline(compute_factors(), 'factor_pipeline')
schedule_function(rebalance,
date_rules.week_start(),
time_rules.market_open(),
calendar=calendars.US_EQUITIES)
set_commission(us_equities=commission.PerShare(cost=0.00075, min_trade_cost=.01))
set_slippage(us_equities=slippage.VolumeShareSlippage(volume_limit=0.0025, price_impact=0.01))
def initialize(context):
context.asset = symbol("AAPL")
# To keep track of whether we invested in the stock or not
context.invested = False
# Explicitly set the commission/slippage to the "old" value until we can
# rebuild example data.
# github.com/quantopian/zipline/blob/master/tests/resources/
# rebuild_example_data#L105
context.set_commission(commission.PerShare(cost=0.0075,
min_trade_cost=1.0))
context.set_slippage(slippage.VolumeShareSlippage())
def initialize(context):
context.i = int(input("Start index?: "))
# Explicitly set the commission/slippage to the "old" value until we can
# rebuild example data.
# github.com/quantopian/zipline/blob/master/tests/resources/
# rebuild_example_data#L105
context.set_commission(commission.PerShare(cost=.005, min_trade_cost=1.0))
context.set_slippage(slippage.VolumeShareSlippage())
context.schedule_function(func=run,
date_rule=date_rules.month_start(),
time_rule=time_rules.market_open(),
half_days=True,
calendar=None)
def initialize(context):
# set_benchmark(symbol('SPY'))
model.init_model()
context.i = 0
context.asset1 = symbol('EBAY')
context.asset2 = symbol('KLAC')
context.model_fee= 1e-3
context.previous_predict_reward=0
context.previous_action=0
context.set_commission(commission.PerShare(cost=0.005, min_trade_cost=1.0))
context.set_slippage(slippage.VolumeShareSlippage())
context.sequence_length=300
context.tb_log_dir='./log/backtest'
context.tensorboard=TensorBoard(log_dir=context.tb_log_dir)
def initialize(context):
# 记录股票代码，通过股票代码获取股票对象
context.asset = symbol('AAPL')
# 定义是否买入股票的标记
context.invested = False
# 设置交易的手续费，股票成交时，手续费按成交金额一定比例收取
# 设置手续费率和最低费用
context.set_commission(commission.PerShare(cost=.0075, min_trade_cost=1.0))
# 设置模拟真实交易的滑价，当实际下单交易时，下单订单将影响市场。买单驱使价格上涨，卖单驱使价格下滑;
# 这通常被称为交易的“价格影响”。价格影响的大小取决于订单与当前交易量相比有多大。
context.set_slippage(slippage.VolumeShareSlippage(volume_limit=0.025, price_impact=0.1))
def initialize(algo, eps=1, window_length=5):
algo.stocks = ['AMD', 'CERN', 'COST', 'DELL', 'GPS', 'INTC', 'MMM']
algo.sids = [algo.symbol(symbol) for symbol in algo.stocks]
algo.m = len(algo.stocks)
algo.price = {}
algo.b_t = np.ones(algo.m) / algo.m
algo.last_desired_port = np.ones(algo.m) / algo.m
algo.eps = eps
algo.init = True
algo.days = 0
algo.window_length = window_length
algo.set_commission(commission.PerShare(cost=0, min_trade_cost=1.0))
algo.set_slippage(slippage.VolumeShareSlippage())
def initialize(context):
attach_pipeline(make_pipeline(), 'my_pipeline')
# Rebalance each day. In daily mode, this is equivalent to putting
# `rebalance` in our handle_data, but in minute mode, it's equivalent to
# running at the start of the day each day.
schedule_function(rebalance, date_rules.every_day())
# Explicitly set the commission/slippage to the "old" value until we can
# rebuild example data.
# github.com/quantopian/zipline/blob/master/tests/resources/
# rebuild_example_data#L105
context.set_commission(commission.PerShare(cost=.0075, min_trade_cost=1.0))
context.set_slippage(slippage.VolumeShareSlippage())
def initialize(context):
context.asset = symbol('000001.SZ')
# Explicitly set the commission/slippage to the "old" value until we can
# rebuild example data.
# github.com/quantopian/zipline/blob/master/tests/resources/
# rebuild_example_data#L105
context.set_commission(commission.PerShare(cost=.0075, min_trade_cost=1.0))
context.set_slippage(slippage.VolumeShareSlippage())
schedule_function(
my_func,
date_rules.every_day(),
time_rules.market_open(minutes=15)
)
def initialize(self, context, stock_ids):
super().initialize(context, stock_ids)
context.portfolio_highest = {}
context.price_highest = {}
context.stock_shares = {}
context.set_commission(
commission.PerShare(cost=.0075, min_trade_cost=1.0))
context.set_slippage(slippage.VolumeShareSlippage())
set_benchmark(context.assets[0])
for asset in context.assets:
context.price_highest[asset] = 0.0
context.portfolio_highest[asset] = 0.0
context.stock_shares[asset] = 0
def initialize(self, eps=1, window_length=5):
self.stocks = STOCKS
self.m = len(self.stocks)
self.price = {}
self.b_t = np.ones(self.m) / self.m
self.last_desired_port = np.ones(self.m) / self.m
self.eps = eps
self.init = True
self.days = 0
self.window_length = window_length
self.add_transform(MovingAverage, 'mavg', ['price'],
window_length=window_length)
no_delay = datetime.timedelta(minutes=0)
slip = slippage.VolumeShareSlippage(volume_limit=0.25,
price_impact=0,
delay=no_delay)
self.set_slippage(slip)
self.set_commission(commission.PerShare(cost=0))
def initialize(context):
# http://money.usnews.com/funds/etfs/rankings/small-cap-funds
context.stocks = [
symbol('SLY'),
symbol('EES'),
symbol('SCHA'),
symbol('IJR'),
symbol('PSCH'),
symbol('VB'),
symbol('VTWO'),
symbol('IWM'),
symbol('SCHC'),
symbol('JKJ')
]
context.m = len(context.stocks)
context.price = {}
context.b_t = np.ones(context.m) / context.m
context.eps = 2.5 #change epsilon here
context.init = False
set_slippage(
slippage.VolumeShareSlippage(volume_limit=0.25, price_impact=0))
set_commission(commission.PerShare(cost=0))
def initialize(context):
context.set_benchmark(None)
context.i = 1
context.assets = list(
map(lambda x: symbol(x), high_cap_company.Symbol.values))
print(context.assets, len(context.assets))
context.model_fee = 5e-3
context.set_commission(commission.PerShare(cost=0.005, min_trade_cost=1.0))
context.set_slippage(slippage.VolumeShareSlippage())
context.bootstrap_sequence_length = 300
context.max_sequence_length = 60
context.tb_log_dir = './log/%s' % back_test_name
context.model_update_time = 30
context.target_profit_multiplier = 1.1
context.model_summaries = None
bundle = bundles.load('quandl')
start_date_str = str(context.get_datetime().date())
initial_history_start_date = bundle.equity_daily_bar_reader.sessions[
bundle.equity_daily_bar_reader.sessions < start_date_str][(
-context.bootstrap_sequence_length - 1)]
initial_history_end_date = bundle.equity_daily_bar_reader.sessions[
bundle.equity_daily_bar_reader.sessions > start_date_str][0]
filterd_assets_index = (np.isnan(
np.sum(bundle.equity_daily_bar_reader.load_raw_arrays(
columns=['close'],
start_date=initial_history_start_date,
end_date=initial_history_end_date,
assets=context.assets),
axis=1)).flatten() == False)
context.assets = list(np.array(context.assets)[filterd_assets_index])
print(context.assets, len(context.assets))
remain_symbols = list(map(lambda x: x.symbol, context.assets))
if not os.path.exists('history_data'):
print('Start to download good history data')
history_data = {}
for s in remain_symbols:
print('downloading', s)
stock = quandl.get_table(
'WIKI/PRICES',
date={'gte': str(initial_history_start_date)},
ticker=s)
stock.index = stock.date
history_data[s] = stock
history_data = pd.Panel(history_data)
history_data.to_pickle('history_data')
context.history_data = generate_stock_features(history_data)
print('Done')
else:
print('history data exist')
history_data = pd.read_pickle('history_data')
context.history_data = generate_stock_features(history_data)
if not os.path.exists('index'):
print('downloading index data')
spy = quandl.get("CHRIS/CME_SP1", authtoken="CTq2aKvtCkPPgR4L_NFs")
gc = quandl.get("CHRIS/CME_GC1", authtoken="CTq2aKvtCkPPgR4L_NFs")
si = quandl.get("CHRIS/CME_SI1", authtoken="CTq2aKvtCkPPgR4L_NFs")
vix = pd.read_csv(
'http://www.cboe.com/publish/scheduledtask/mktdata/datahouse/vixcurrent.csv'
)
vix.columns = vix.iloc[0]
vix = vix[1:]
vix.index = pd.DatetimeIndex(vix.Date)
vix = vix.drop('Date', axis=1)
vix = vix.astype(np.float64)
vix.columns = ['Open', 'High', 'Low', 'Last']
index_data = pd.Panel({'vix': vix, 'gc': gc, 'si': si, 'spy': spy})
index_data.to_pickle('index')
index_data = index_data[:, str(initial_history_start_date):, :]
context.index_data = generate_index_features(index_data)
else:
print('index data exist')
index_data = pd.read_pickle('index')
index_data = index_data[:, str(initial_history_start_date):, :]
context.index_data = generate_index_features(
index_data)[:, context.history_data.major_axis[0]:, :]
if not os.path.exists('trading_content'):
sys.exit(1)
else:
news_vec = pd.read_csv('trading_content')
news_vec.index = news_vec.date
news_vec = news_vec.drop('date', axis=1)
news_vec = context.history_data[:, :,
'return_rate'].join(news_vec).drop(
context.history_data.items,
axis=1).fillna(0)
context.news_vec = news_vec
assert context.history_data.major_axis[0] == context.index_data.major_axis[
0]
feature_network_topology = {
'equity_network': {
'feature_map_number': len(context.assets),
'feature_number': context.history_data.shape[2],
'input_name': 'equity',
'dense': {
'n_units': [128, 64],
'act': [tf.nn.tanh] * 2,
},
'rnn': {
'n_units': [32, 1],
'act': [tf.nn.tanh, None],
'attention_length': 10
},
'keep_output': True
},
'index_network': {
'feature_map_number': len(context.index_data.items),
'feature_number': context.index_data.shape[2],
'input_name': 'index',
'dense': {
'n_units': [128, 64],
'act': [tf.nn.tanh] * 2,
},
'rnn': {
'n_units': [32, 16],
'act': [tf.nn.tanh, tf.nn.tanh],
'attention_length': 10
},
'keep_output': False
},
'weight_network': {
'feature_map_number': 1,
'feature_number': len(context.assets) + 1,
'input_name': 'weight',
'dense': {
'n_units': [32, 16],
'act': [tf.nn.tanh] * 2,
},
'rnn': {
'n_units': [16, 8],
'act': [tf.nn.tanh, tf.nn.tanh],
'attention_length': 10
},
'keep_output': False
},
'return_network': {
'feature_map_number': 1,
'feature_number': 1,
'input_name': 'return',
'dense': {
'n_units': [8, 4],
'act': [tf.nn.tanh] * 2,
},
'rnn': {
'n_units': [4, 2],
'act': [tf.nn.tanh, tf.nn.tanh],
'attention_length': 10
},
'keep_output': False
},
'news_network': {
'feature_map_number': 1,
'feature_number': 100,
'input_name': 'return',
'dense': {
'n_units': [128, 64],
'act': [tf.nn.tanh] * 2,
},
'rnn': {
'n_units': [32, 16],
'act': [tf.nn.tanh, tf.nn.tanh],
'attention_length': 10
},
'keep_output': False
}
}
context.model = DRL_Portfolio(
asset_number=len(context.assets),
feature_network_topology=feature_network_topology,
action_network_layers=[32, 16],
object_function='reward')
context.real_return = []
context.history_weight = []
context.model.init_model()
context.tensorboard = TensorBoard(log_dir=context.tb_log_dir,
session=context.model.get_session())
def initialize(context):
context.i = 0
context.assets = list(
map(lambda x: symbol(x), high_cap_company.Symbol.values))
print(context.assets, len(context.assets))
context.model_fee = 1e-3
context.previous_predict_reward = 0
context.previous_action = 0
context.set_commission(commission.PerShare(cost=0.005, min_trade_cost=1.0))
context.set_slippage(slippage.VolumeShareSlippage())
context.sequence_length = 300
context.tb_log_dir = './log/backtest'
context.tensorboard = TensorBoard(log_dir=context.tb_log_dir)
bundle = bundles.load('quandl')
start_date_str = str(context.get_datetime().date())
initial_history_start_date = bundle.equity_daily_bar_reader.sessions[
bundle.equity_daily_bar_reader.sessions < start_date_str][(
-context.sequence_length - 1)]
initial_history_end_date = bundle.equity_daily_bar_reader.sessions[
bundle.equity_daily_bar_reader.sessions > start_date_str][0]
filterd_assets_index = (np.isnan(
np.sum(bundle.equity_daily_bar_reader.load_raw_arrays(
columns=['close'],
start_date=initial_history_start_date,
end_date=initial_history_end_date,
assets=context.assets),
axis=1)).flatten() == False)
context.assets = list(np.array(context.assets)[filterd_assets_index])
print(context.assets, len(context.assets))
remain_symbols = list(map(lambda x: x.symbol, context.assets))
if not os.path.exists('history_data'):
print('Start to download good history data')
history_data = {}
for s in remain_symbols:
print('downloading', s)
stock = quandl.get_table(
'WIKI/PRICES',
date={'gte': str(initial_history_start_date)},
ticker=s)
stock.index = stock.date
history_data[s] = stock
history_data = pd.Panel(history_data)
history_data = history_data.transpose(2, 1, 0)
history_data.to_pickle('history_data')
context.history_data = history_data
print('Done')
else:
print('history data exist')
history_data = pd.read_pickle('history_data')
context.history_data = history_data
if not os.path.exists('vix.csv'):
print('Start to download VIX index')
vix = quandl.get("CHRIS/CBOE_VX1", authtoken="CTq2aKvtCkPPgR4L_NFs")
vix = vix[str(initial_history_start_date):]
vix.to_csv('vix.csv')
if not os.path.exists('si.csv'):
print('Start to download silver price')
si = quandl.get("CHRIS/CME_SI1", authtoken="CTq2aKvtCkPPgR4L_NFs")
si.to_csv('si.csv')
if not os.path.exists('gc.csv'):
print('Start to download gold price')
gc = quandl.get("CHRIS/CME_GC1", authtoken="CTq2aKvtCkPPgR4L_NFs")
gc.to_csv('gc.csv')
if not os.path.exists('spy.csv'):
print('Start to download SP500 Index')
spy = quandl.get("CHRIS/CME_SP1", authtoken="CTq2aKvtCkPPgR4L_NFs")
spy.to_csv('spy.csv')
context.model = DRL_Portfolio(feature_number=len(context.assets) * 8,
asset_number=len(context.assets) + 1,
object_function='sortino')
context.model.init_model()
def initialize(context):
context.i = 0
context.assets = list(map(lambda x: symbol(x), high_cap_company.Symbol.values))
print(context.assets, len(context.assets))
context.model_fee = 1e-3
context.previous_predict_reward = 0
context.previous_action = 0
context.set_commission(commission.PerShare(cost=0.005, min_trade_cost=1.0))
context.set_slippage(slippage.VolumeShareSlippage())
context.bootstrap_sequence_length = 300
context.max_sequence_length = 500
context.tb_log_dir = './log/%s' % back_test_name
context.tensorboard = TensorBoard(log_dir=context.tb_log_dir)
context.target_profit_multiplier = 1.1
bundle = bundles.load('quandl')
start_date_str = str(context.get_datetime().date())
initial_history_start_date = bundle.equity_daily_bar_reader.sessions[bundle.equity_daily_bar_reader.sessions < start_date_str][(-context.bootstrap_sequence_length - 1)]
initial_history_end_date = bundle.equity_daily_bar_reader.sessions[bundle.equity_daily_bar_reader.sessions > start_date_str][0]
filterd_assets_index = (np.isnan(np.sum(bundle.equity_daily_bar_reader.load_raw_arrays(columns=['close'], start_date=initial_history_start_date, end_date=initial_history_end_date, assets=context.assets), axis=1)).flatten() == False)
context.assets = list(np.array(context.assets)[filterd_assets_index])
print(context.assets, len(context.assets))
remain_symbols = list(map(lambda x: x.symbol, context.assets))
if not os.path.exists('history_data'):
print('Start to download good history data')
history_data = {}
for s in remain_symbols:
print('downloading', s)
stock = quandl.get_table('WIKI/PRICES', date={'gte': str(initial_history_start_date)}, ticker=s)
stock.index = stock.date
history_data[s] = stock
history_data = pd.Panel(history_data)
history_data = history_data.transpose(2, 1, 0)
history_data.to_pickle('history_data')
context.history_data = history_data
print('Done')
else:
print('history data exist')
history_data = pd.read_pickle('history_data')
context.history_data = history_data
if not os.path.exists('trading_content'):
sys.exit(1)
else:
news_vec = pd.read_csv('trading_content')
news_vec.index = news_vec.date
news_vec = news_vec.drop('date', axis=1)
context.news_vec = news_vec
if not os.path.exists('index'):
print('downloading index data')
spy = quandl.get("CHRIS/CME_SP1", authtoken="CTq2aKvtCkPPgR4L_NFs")
gc = quandl.get("CHRIS/CME_GC1", authtoken="CTq2aKvtCkPPgR4L_NFs")
si = quandl.get("CHRIS/CME_SI1", authtoken="CTq2aKvtCkPPgR4L_NFs")
vix = pd.read_csv('http://www.cboe.com/publish/scheduledtask/mktdata/datahouse/vixcurrent.csv')
vix.columns = vix.iloc[0]
vix = vix[1:]
vix.index = pd.DatetimeIndex(vix.Date)
vix = vix.drop('Date', axis=1)
vix = vix.astype(np.float64)
vix.columns = ['Open', 'High', 'Low', 'Last']
index_data = pd.Panel({'vix': vix, 'gc': gc, 'si': si, 'spy': spy})
index_data = index_data.transpose(2, 1, 0)
index_data.to_pickle('index')
context.index_data = index_data['Last', str(initial_history_start_date):]
else:
print('index data exist')
index_data = pd.read_pickle('index')
context.index_data = index_data['Last', str(initial_history_start_date):]
context.model = DRL_Portfolio(feature_number=len(context.assets) * 8 + 100 + context.index_data.columns.shape[0] * 7, asset_number=len(context.assets) + 1, object_function='reward')
context.model.init_model()
def initialize(context):
context.asset = symbol('BITMAP')
context.invested = False
context.set_commission(commission.PerShare(cost=.0075, min_trade_cost=1.0))
context.set_slippage(slippage.VolumeShareSlippage())
def initialize(self):
self.set_commission(commission.PerShare(cost=self.transaction_cost, min_trade_cost=1.0))
self.set_slippage(slippage.VolumeShareSlippage())
def initialize( self ):
# Adjust slippage
self.set_slippage( slippage.VolumeShareSlippage( volume_limit = 1.0, price_impact = 0.01 ) )
# Trade on boundary of first trading day of MONTH or set DAYS
# DAYS|MONTH
# self.boundaryTrade = 'DAYS'
# self.boundaryDays = 21
self.boundaryTrade = 'MONTH'
# Set Performance vs. Volatility factors (7.0, 3.0 from Grossman GMRE
self.factorPerformance = 0.7
self.factorVolatility = 0.3
# Period Volatility and Performance period in DAYS
self.metricPeriod = 63 # 3 months LOOKBACK
self.periodVolatility = 21 # Volatility period. Chose a MULTIPLE of metricPeriod
# To prevent going 'negative' on cash account set stop, limit and price factor >= stop
# self.orderBuyLimits = False
# self.orderSellLimits = False
# #self.priceBuyStop = None
# #self.priceBuyLimit = None
# #self.priceSellStop = None
# #self.priceSellLimit = None
# self.priceBuyFactor = 3.03 # Buffering since buys and sells DON'T occur on the same day.
# Re-enact pricing from original Quast code
self.orderBuyLimits = False
self.orderSellLimits = False
self.priceBuyFactor = 0.0
# Factor commission cost
# self.set_commission(commission.PerShare(cost=0.03))
# self.set_commission(commission.PerTrade(cost=15.00))
# Set the basket of stocks
self.basket = {
12915: 'MDY', # MDY (SPDR S&P MIDCAP 400)
21769: 'IEV', # IEV (ISHARES EUROPE ETF)
24705: 'EEM', # EEM (ISHARES MSCI EMERGING MARKETS)
23134: 'ILF', # ILF (ISHARES LATIN AMERICA 40)
23118: 'EPP', # EPP (ISHARES MSCI PACIFIC EX JAPAN)
22887: 'EDV', # EDV (VANGUARD EXTENDED DURATION TREASURY)
40513: 'ZIV', # ZIV (VelocityShares Inverse VIX Medium-Term)
23911: 'SHY', # SHY (iShares 1-3 Year Treasury Bond ETF)
}
self.logWarn = True
self.logBuy = True
self.logSell = True
self.logHold = True
self.logRank = False
self.logDebug = False
# SHOULDN'T NEED TO MODIFY REMAINING VARIABLES
self.accumulateData = accumulateData( window_length = self.metricPeriod )
# Keep track of the current period
self.cashStart = None
self.dateStart = None
self.currentDayNum = None
self.currentMonth = None
self.currentStock = None
self.nextStock = None
self.oidBuy = None
self.oidSell = None
self.period_start_portfolio_value = None
self.buyCount = 0
self.sellCount = 0
