def __get_time4candles(self, n, anchor_point, roll=True):
'''
This private function takes a a number of candles
and returns a Datetime corresponding to
this number of candles
Parameters
----------
n : int
Number of candles
anchor_point : datetime
Datetime used as the anchor (end point from which it will go back 'n' candles) for calculation
roll : boolean
if True, then if will try to go back in time in order to get exactly 'n' number of candles.
Default: True
Returns
-------
Datetime.datetime
'''
delta_from_start = None
delta_one = None
if self.timeframe == "D":
delta_from_start = datetime.timedelta(hours=24 * n)
delta_one = datetime.timedelta(hours=24)
else:
fgran = self.timeframe.replace('H', '')
delta_from_start = datetime.timedelta(hours=int(fgran) * n)
delta_one = datetime.timedelta(hours=int(fgran))
# calculate the cutoff for the first threshold using the number of candles
oanda = OandaAPI(
url=config.OANDA_API['url'],
instrument=self.pair,
granularity=self.timeframe,
alignmentTimezone=config.OANDA_API['alignmentTimezone'],
dailyAlignment=config.OANDA_API['dailyAlignment'])
start = anchor_point - delta_from_start
if roll is True:
if start < config.START_HIST[self.pair]:
#return the first candle in the record if start goes back before the start of the record
return config.START_HIST[self.pair]
else:
end = anchor_point.isoformat()
oanda.run(start=start.isoformat(), end=end, roll=True)
candle_list = oanda.fetch_candleset()
# if len of candle_list is below n then go back one candle at a time
while len(candle_list) < n:
start = start - delta_one
oanda.run(start=start.isoformat(), end=end, roll=True)
candle_list = oanda.fetch_candleset()
return start
else:
return start
def __initclist(self):
'''
Private function to initialize the CandleList object that goes from self.start
to self.period
This will set the self.clist_period class attribute
'''
delta_period = periodToDelta(self.period, self.timeframe)
delta_1 = periodToDelta(1, self.timeframe)
start = self.start - delta_period # get the start datetime for this CandleList period
end = self.start + delta_1 # increase self.start by one candle to include self.start
oanda = OandaAPI(
url=config.OANDA_API['url'],
instrument=self.pair,
granularity=self.timeframe,
alignmentTimezone=config.OANDA_API['alignmentTimezone'],
dailyAlignment=config.OANDA_API['dailyAlignment'])
oanda.run(start=start.isoformat(), end=end.isoformat(), roll=True)
candle_list = oanda.fetch_candleset(vol_cutoff=0)
cl = CandleList(candle_list,
self.pair,
granularity=self.timeframe,
id=self.id,
type=self.type)
self.clist_period = cl
def fetch_candlelist(self):
'''
This function returns a CandleList object for this Trade
Returns
-------
A CandleList object
'''
oanda = OandaAPI(
url=config.OANDA_API['url'],
instrument=self.pair,
granularity=self.timeframe,
alignmentTimezone=config.OANDA_API['alignmentTimezone'],
start=datetime.datetime.strptime(self.start,
'%Y-%m-%dT%H:%M:%S').isoformat(),
dailyAlignment=config.OANDA_API['dailyAlignment'],
end=datetime.datetime.strptime(self.end,
'%Y-%m-%dT%H:%M:%S').isoformat())
candle_list = oanda.fetch_candleset()
cl = CandleList(candle_list, type=self.type)
return cl
def __init_clist_period(self):
'''
Private function to initialise self.clist_period class attribute
This function process the candlelist going from self.start-self.period to
self.start
'''
warnings.warn("[INFO] Run __init_clist_period")
delta_period = None
delta_1 = None
if self.timeframe == "D":
delta_period = datetime.timedelta(hours=24 * self.period)
delta_1 = datetime.timedelta(hours=24)
else:
fgran = self.timeframe.replace('H', '')
delta_period = datetime.timedelta(hours=int(fgran) * self.period)
delta_1 = datetime.timedelta(hours=int(fgran))
start = self.start - delta_period
end = self.start + delta_1
oanda = OandaAPI(
url=config.OANDA_API['url'],
instrument=self.pair,
granularity=self.timeframe,
alignmentTimezone=config.OANDA_API['alignmentTimezone'],
dailyAlignment=config.OANDA_API['dailyAlignment'])
oanda.run(start=start.isoformat(), end=end.isoformat(), roll=True)
candle_list = oanda.fetch_candleset(vol_cutoff=0)
cl = CandleList(candle_list,
self.pair,
granularity=self.timeframe,
id=self.id)
warnings.warn("[INFO] Run cl.calc_rsi")
cl.calc_rsi()
warnings.warn("[INFO] Done cl.calc_rsi")
self.clist_period = cl
def get_cross_time(self, candle, granularity='M30'):
'''
This function is used get the time that the candle
crosses (go through) HArea
Parameters
----------
candle : Candle object that crosses the HArea
granularity : To what granularity we should descend
Returns
------
datetime object with crossing time.
n.a. if crossing time could not retrieved. This can happens
when there is an artifactual jump in the Oanda data
'''
if candle.lowAsk <= self.price <= candle.highAsk:
delta = None
if self.granularity == "D":
delta = timedelta(hours=24)
else:
fgran = self.granularity.replace('H', '')
delta = timedelta(hours=int(fgran))
cstart = candle.time
cend = cstart + delta
oanda = OandaAPI(
url=config.OANDA_API['url'],
instrument=self.instrument,
granularity=granularity,
dailyAlignment=config.OANDA_API['dailyAlignment'],
alignmentTimezone=config.OANDA_API['alignmentTimezone'])
oanda.run(start=cstart.isoformat(),
end=cend.isoformat(),
roll=True)
candle_list = oanda.fetch_candleset()
for c in candle_list:
if c.lowAsk <= self.price <= c.highAsk:
return c.time
else:
return 'n.a.'
def cl_object():
'''Returns CandleList object'''
oanda = OandaAPI(url='https://api-fxtrade.oanda.com/v1/candles?',
instrument='AUD_USD',
granularity='D',
alignmentTimezone='Europe/London',
dailyAlignment=22)
oanda.run(start='2018-01-25T22:00:00',
end='2018-10-12T22:00:00',
roll=True)
candle_list = oanda.fetch_candleset()
cl = CandleList(candle_list, instrument='AUD_USD', type='long')
return cl
def __init_clist_trend(self):
'''
Private function to initialise self.clist_trend class attribute
This function process the candlelist going from self.trend_i to self.start
'''
warnings.warn("[INFO] Run __init_clist_trend")
# if trend_i is not defined then calculate it
if hasattr(self, 'trend_i'):
self.trend_i = datetime.datetime.strptime(self.trend_i,
'%Y-%m-%d %H:%M:%S')
else:
self.calc_itrend()
self.__init_clist_trend()
# checking for feats in trend before 1st bounce
oanda = OandaAPI(
url=config.OANDA_API['url'],
instrument=self.pair,
granularity=self.timeframe,
alignmentTimezone=config.OANDA_API['alignmentTimezone'],
dailyAlignment=config.OANDA_API['dailyAlignment'])
oanda.run(start=self.trend_i.isoformat(),
end=self.start.isoformat(),
roll=True)
candle_list = oanda.fetch_candleset(vol_cutoff=0)
cl = CandleList(candle_list,
instrument=self.pair,
granularity=self.timeframe,
id=self.id)
warnings.warn("[INFO] Run cl.calc_rsi")
cl.calc_rsi()
warnings.warn("[INFO] Done cl.calc_rsi")
self.clist_trend = cl
def pl_object2():
'''Returns PivotList object'''
oanda = OandaAPI(url='https://api-fxtrade.oanda.com/v1/candles?',
instrument='AUD_USD',
granularity='D',
alignmentTimezone='Europe/London',
dailyAlignment=22)
oanda.run(start='2016-01-15T22:00:00',
end='2016-08-17T22:00:00',
roll=True)
candle_list = oanda.fetch_candleset()
cl = CandleList(candle_list, instrument='AUD_USD', type='long')
pl = cl.get_pivotlist(outfile='test.png', th_up=0.02, th_down=-0.02)
return pl
def set_valley(self):
'''
Function to calculate the length of the valley
between bounce_1st & bounce_2nd
Returns
-------
It will set the 'valley' attribute of the class
'''
oanda = OandaAPI(
url=config.OANDA_API['url'],
instrument=self.pair,
granularity=self.timeframe,
alignmentTimezone=config.OANDA_API['alignmentTimezone'],
dailyAlignment=config.OANDA_API['dailyAlignment'])
oanda.run(start=self.bounce_2nd.time.isoformat(),
end=self.bounce_1st.time.isoformat())
candle_list = oanda.fetch_candleset(vol_cutoff=0)
self.valley = len(candle_list)
from oanda_api import OandaAPI
import logging
logging.basicConfig(level=logging.INFO)
logging.info("Program started")
#loanda=OandaAPI(url='https://api-fxtrade.oanda.com/v1/candles?',instrument='EUR_USD',granularity='H8',alignmentTimezone='Europe/London',dailyAlignment=22,start='2016-09-13T06:00:00',count=40)
oanda=OandaAPI(url='https://api-fxtrade.oanda.com/v1/candles?',instrument='EUR_USD',granularity='H8',alignmentTimezone='Europe/London',dailyAlignment=22,start='2016-10-24T22:00:00',count=1)
print(oanda.print_url())
candlelist=oanda.fetch_candleset()
for c in candlelist:
c.set_candle_features()
c.set_candle_formation()
print("%s %s" % (c.representation,c.time))
logging.info("Done!")
def calc_rsi(self, period=2000, rsi_period=14):
'''
Calculate the RSI for a certain candle list
Parameters
----------
period : int
Number of candles before this CandleList start for which close price data will be fetched.
The larger the number of candles the more accurate the ewm calculation will be, as the exponential
moving average calculated for each of the windows (of size=rsi_period) will be
directly affected by the previous windows in the series. Default=2000
rsi_period : int
Number of candles used for calculating the RSI. Default=14
Returns
-------
Nothing
'''
start_time=self.clist[0].time
end_time=self.clist[-1].time
delta_period = None
if self.granularity == "D":
delta_period = datetime.timedelta(hours=24 * period)
else:
fgran = self.granularity.replace('H', '')
delta_period = datetime.timedelta(hours=int(fgran) * period)
start_calc_time = start_time - delta_period
#fetch candle set from start_calc_time
oanda = OandaAPI(url=config.OANDA_API['url'],
instrument=self.instrument,
granularity=self.granularity,
alignmentTimezone=config.OANDA_API['alignmentTimezone'],
dailyAlignment=config.OANDA_API['dailyAlignment'])
'''
Get candlelist from start_calc_time to (start_time-1)
This 2-step API call is necessary in order to avoid
maximum number of candles errors
'''
oanda.run(start=start_calc_time.isoformat(),
end=start_time.isoformat(),
roll=True)
cl1 = oanda.fetch_candleset()
'''Get candlelist from start_time to end_time'''
oanda.run(start=start_time.isoformat(),
end=end_time.isoformat(),
roll=True)
cl2 = oanda.fetch_candleset()
if cl1[-1].time == cl2[0].time:
del cl1[-1]
candle_list = cl1 + cl2
series=[]
series = [c.closeAsk for c in candle_list]
df = pd.DataFrame({'close': series})
chg = df['close'].diff(1)
gain = chg.mask(chg < 0, 0)
loss = chg.mask(chg > 0, 0)
avg_gain = gain.ewm(com=rsi_period - 1, min_periods=rsi_period).mean()
avg_loss = loss.ewm(com=rsi_period - 1, min_periods=rsi_period).mean()
rs = abs(avg_gain / avg_loss)
rsi = 100 - (100 / (1 + rs))
rsi4cl=rsi[-len(self.clist):]
# set rsi attribute in each candle of the CandleList
ix=0
for c,v in zip(self.clist,rsi4cl):
self.clist[ix].rsi=v
ix+=1
def run_trade(self):
'''
Run the trade until conclusion from a start date
'''
print("[INFO] Run run_trade with id: {0}".format(self.id))
entry = HArea(price=self.entry,
pips=1,
instrument=self.pair,
granularity=self.timeframe)
SL = HArea(price=self.SL,
pips=1,
instrument=self.pair,
granularity=self.timeframe)
TP = HArea(price=self.TP,
pips=1,
instrument=self.pair,
granularity=self.timeframe)
period = None
if self.timeframe == "D":
period = 24
else:
period = int(self.timeframe.replace('H', ''))
# generate a range of dates starting at self.start and ending numperiods later in order to assess the outcome
# of trade and also the entry time
self.start = datetime.datetime.strptime(str(self.start),
'%Y-%m-%d %H:%M:%S')
numperiods = 300
date_list = [
datetime.datetime.strptime(str(self.start.isoformat()),
'%Y-%m-%dT%H:%M:%S') +
datetime.timedelta(hours=x * period) for x in range(0, numperiods)
]
entered = False
for d in date_list:
oanda = OandaAPI(
url=config.OANDA_API['url'],
instrument=self.pair,
granularity=self.timeframe,
dailyAlignment=config.OANDA_API['dailyAlignment'],
alignmentTimezone=config.OANDA_API['alignmentTimezone'])
oanda.run(start=d.isoformat(), count=1, roll=True)
cl = oanda.fetch_candleset()[0]
if entered is False:
entry_time = entry.get_cross_time(candle=cl)
warnings.warn("\t[INFO] Trade entered")
if entry_time != 'n.a.':
self.entry_time = entry_time.isoformat()
else:
warnings.warn(
"No entry time was identified for this trade")
entry_time = self.start
self.entry_time = entry_time
if entry_time is not None and entry_time != 'n.a.':
entered = True
if entered is True:
failure_time = SL.get_cross_time(candle=cl)
if failure_time is not None and failure_time != 'n.a.':
self.outcome = 'failure'
self.end = failure_time
self.pips = float(
calculate_pips(self.pair,
abs(self.SL - self.entry))) * -1
warnings.warn("\t[INFO] S/L was hit")
break
if entered is True:
success_time = TP.get_cross_time(candle=cl)
if success_time is not None and success_time != 'n.a.':
self.outcome = 'success'
warnings.warn("\t[INFO] T/P was hit")
self.end = success_time
self.pips = float(
calculate_pips(self.pair, abs(self.TP - self.entry)))
break
assert getattr(self, 'outcome')
warnings.warn("[INFO] Done run_trade")