Hi, I see some code in strategy as follows:
def __init__(self):
# To control operation entries
self.orderid = None
# Create SMA on 2nd data
sma = btind.MovAv.SMA(self.data1, period=self.p.period)
# Create a CrossOver Signal from close an moving average
self.signal = btind.CrossOver(self.data1.close, sma)
I wonder why the sma isn't prefixed by self. while signal is. Is there any guidance tell us when to use self or not?
Got it. I misunderstood something stupidly.
@ab_trader thank you for your reply.
Suppose the data is as follows：
2019-01-01, ...
2019-02-15, ...
At first bar, that is 2019-1-1, I issue an order, then it will be executed at the second bar, that is 2019-2-15. Now, I want the order to expire on 2019-1-10. How to do it?
I think if I can set a valid day for the market order, the problem can be solved easily.
My fault. Thank you.
Hi, backtraders. In Chinese futures market, for the same contract, you can hold long and short positions simultaneously. and we have four commands to submit an order: open_long, close_long, open_short, close_short.
I wonder if it's possible to customize bt to support Chinese future market？
Hello, backtrader. I have a question on replay. pls see the following.
class SmaCross(bt.Strategy):
def __init__(self):
self.SMA0 = bt.ind.MovingAverageSimple(self.data0, period=5)
self.SMA1 = bt.ind.MovingAverageSimple(self.data1, period=5)
def next(self):
pass
if self.data0.datetime.datetime() == datetime.datetime(2010, 1, 1, 12,
19) or self.data0.datetime.datetime() == datetime.datetime(
2010, 1, 1, 12, 18):
print(self.data0.datetime.datetime(), 'd1_sma', self.SMA1[0], 'd1_sma[-1]', self.SMA1[-1])
##########################
# main
#########################
cerebro = bt.Cerebro()
data = bt.feeds.GenericCSVData(
dataname='./candles00.csv',
timeframe=bt.TimeFrame.Minutes,
) # 1 minute bar
cerebro.adddata(data, name='1min')
cerebro.replaydata(data, timeframe=bt.TimeFrame.Minutes, compression=60, name='60m') # 60 minute bar
cerebro.addstrategy(SmaCross)
cerebro.broker.setcash(10000)
results = cerebro.run()
and the following is the output.
I expect that sma[-1] for data1 at 12:19 should be equal to sma[0] at 12:18, but they are not. I wonder how to make them equal.
I know it is created and start as a thread by streaming_prices. But I don't know who whill call back it.
Hi, friends. I am studying OandaV20Store, there is method:
def _t_streaming_prices(self, dataname, q):
'''Callback method for streaming prices'''
try:
response = self.oapi_stream.pricing.stream(
self.p.account,
instruments=dataname,
)
# process response
for msg_type, msg in response.parts():
if msg_type == 'pricing.ClientPrice':
# put price into queue as dict
q.put(msg.dict())
The comment says it is a Callback method for streaming prices. I guess every time a new data come, then the method will be called. I wonder which method trigger this method? could someone gives me a hint?
@vladisld Thank you. Below is the load method, could you point me which line of code will fetch bar data in the queue?
def _load(self):
if self._state == self._ST_OVER:
return False
while True:
if self._state == self._ST_LIVE:
try:
msg = (self._storedmsg.pop(None, None) or
self.qlive.get(timeout=self._qcheck))
except queue.Empty:
return None # indicate timeout situation
if msg is None: # Conn broken during historical/backfilling
self.put_notification(self.CONNBROKEN)
# Try to reconnect
if not self.p.reconnect or self._reconns == 0:
# Can no longer reconnect
self.put_notification(self.DISCONNECTED)
self._state = self._ST_OVER
return False # failed
self._reconns -= 1
self._st_start(instart=False, tmout=self.p.reconntimeout)
continue
if 'code' in msg:
self.put_notification(self.CONNBROKEN)
code = msg['code']
if code not in [599, 598, 596]:
self.put_notification(self.DISCONNECTED)
self._state = self._ST_OVER
return False # failed
if not self.p.reconnect or self._reconns == 0:
# Can no longer reconnect
self.put_notification(self.DISCONNECTED)
self._state = self._ST_OVER
return False # failed
# Can reconnect
self._reconns -= 1
self._st_start(instart=False, tmout=self.p.reconntimeout)
continue
self._reconns = self.p.reconnections
# Process the message according to expected return type
if not self._statelivereconn:
if self._laststatus != self.LIVE:
if self.qlive.qsize() <= 1: # very short live queue
self.put_notification(self.LIVE)
ret = self._load_tick(msg)
if ret:
return True
# could not load bar ... go and get new one
continue
# Fall through to processing reconnect - try to backfill
self._storedmsg[None] = msg # keep the msg
# else do a backfill
if self._laststatus != self.DELAYED:
self.put_notification(self.DELAYED)
dtend = None
if len(self) > 1:
# len == 1 ... forwarded for the 1st time
dtbegin = self.datetime.datetime(-1)
elif self.fromdate > float('-inf'):
dtbegin = num2date(self.fromdate)
else: # 1st bar and no begin set
# passing None to fetch max possible in 1 request
dtbegin = None
dtend = datetime.utcfromtimestamp(int(msg['time']) / 10 ** 6)
self.qhist = self.o.candles(
self.p.dataname, dtbegin, dtend,
self._timeframe, self._compression,
candleFormat=self._candleFormat,
includeFirst=self.p.includeFirst)
self._state = self._ST_HISTORBACK
self._statelivereconn = False # no longer in live
continue
elif self._state == self._ST_HISTORBACK:
msg = self.qhist.get()
if msg is None: # Conn broken during historical/backfilling
# Situation not managed. Simply bail out
self.put_notification(self.DISCONNECTED)
self._state = self._ST_OVER
return False # error management cancelled the queue
elif 'code' in msg: # Error
self.put_notification(self.NOTSUBSCRIBED)
self.put_notification(self.DISCONNECTED)
self._state = self._ST_OVER
return False
if msg:
if self._load_history(msg):
return True # loading worked
continue # not loaded ... date may have been seen
else:
# End of histdata
if self.p.historical: # only historical
self.put_notification(self.DISCONNECTED)
self._state = self._ST_OVER
return False # end of historical
# Live is also wished - go for it
self._state = self._ST_LIVE
continue
elif self._state == self._ST_FROM:
if not self.p.backfill_from.next():
# additional data source is consumed
self._state = self._ST_START
continue
# copy lines of the same name
for alias in self.lines.getlinealiases():
lsrc = getattr(self.p.backfill_from.lines, alias)
ldst = getattr(self.lines, alias)
ldst[0] = lsrc[0]
return True
elif self._state == self._ST_START:
if not self._st_start(instart=False):
self._state = self._ST_OVER
return False
Dear all, I am studying Oandadata feed. For now, I have known how Oandadata get data from remote point. Please see this.
The data is put into a queue q.
Now, I wonder how backtrader fetches the data in the queue and triggers next method in strategy.
Any help is appreciated.
Dear all, I wonder what's the difference between PandasDirectData and PandasData? According to the document, both use dataframe.
@rajanprabu is timeframe informative too？ That means it will not influence any other part of the strategy?
Hi, in this post https://community.backtrader.com/topic/309/hourly-bars/1
@Serp said:
thank you very much. English is not my native language and I thought the sentence:
"Informative unless Data Resampling/Replaying is used" would mean, that it is just for my personal informative use (like I could print that it is "daily", but it has no other purpose).
Thanks, so to sum it up (for other people reading here) the solution was:
timeframe = bt.TimeFrame.Minutes,
compression = 60,
And @backtrader relpied: You read it right and has been corrected.
My question is if both "timeframe" and "compression" are informative or only "compression" is informative in data feed?
Hope someone could help.
@crunchypickle do you mean you can use pyfolio now?