Sorry, more about missing attribute 'high', and definition in init not recognised in next

andydoc1

So i have a script which is giving me headaches, mostly with "'Lines_LineSeries_LineSeriesStub' object has no attribute 'high'", but also with failure to recognise a signal from init, in next. The df1 dataframe is as follows (head):

open	close	high	low
date				
2018-12-18 00:05:00	1.13497	1.13518	1.13519	1.13495
2018-12-18 00:10:00	1.13519	1.13533	1.13533	1.13508
2018-12-18 00:15:00	1.13533	1.13527	1.13542	1.13491
2018-12-18 00:20:00	1.13528	1.13517	1.13529	1.13517
2018-12-18 00:25:00	1.13516	1.13508	1.13516	1.13501
2018-12-18 00:30:00	1.13508	1.13495	1.13519	1.13495
2018-12-18 00:35:00	1.13493	1.13462	1.13493	1.13444
2018-12-18 00:40:00	1.13462	1.13471	1.13472	1.13444
2018-12-18 00:45:00	1.13470	1.13478	1.13483	1.13462
2018-12-18 00:50:00	1.13478	1.13493	1.13495	1.13465

The script thus:

class TestStrategy(bt.Strategy):
  
    def log(self, txt, dt=None):
        ''' Logging function for this strategy'''
        dt = dt or self.datas[0].datetime.datetime(0)
        print('%s, %s' % (dt.isoformat(sep=' '), txt))

    def __init__(self):
        # Keep a reference to the "close" line in the data[0] dataseries
        self.dataclose = self.datas[0].close
        self.order = None
        
        trending = bt.indicators.ADX(self.data[0], period=14)>25
        directionup = bt.indicators.PlusDI(self.data[0], period=14) - bt.indicators.MinusDI(self.data[0], period=14)>0
        directiondn = bt.indicators.PlusDI(self.data[0], period=14) - bt.indicators.MinusDI(self.data[0], period=14)<0
        buy_sig = bt.And(trending, directionup)
        sell_sig = bt.And(trending, directiondn)

    def notify_order(self,order):
      if order.status in [order.Submitted, order.Accepted]:
        return
      
      if order.status == order.Completed:
        if order.isbuy():
          self.log('Buy Executed @ %.5f' % order.executed.price)
        elif order.issell():
          self.log('Sell Executed, %.5f' % order.executed.price)

        self.bar_executed = len(self)

      self.order = None

    def next(self):
        # Simply log the closing price of the series from the reference
        self.log('Close, %.5f' % self.dataclose[0])
        print(self.position)

#following line added for debugging
        print(trending)
  
        if buy_sig:
          self.log('BUY CREATE, %.5f' % self.dataclose[0])
          self.order = self.buy()
        elif sell_sig:
          self.log('SELL CREATE, %.5f' % self.dataclose[0])
          self.order = self.sell() 

cerebro = bt.Cerebro(exactbars = False)
cerebro.broker.set_cash(50000)
df1 = df.iloc[:, 0:4]
data = bt.feeds.PandasData(dataname=df1
                           ,fromdate=datetime.datetime(2018, 12, 18),
                            # Do not pass values before this date
                            todate=datetime.datetime(2019, 3, 1)
                            # Do not pass values after this date
                          )
cerebro.adddata(data)
cerebro.addstrategy(TestStrategy)
cerebro.addsizer(bt.sizers.FixedSize, stake=500)
print('Starting Portfolio Value: %.2f' % cerebro.broker.getvalue())

cerebro.run()

print('Final Portfolio Value: %.2f' % cerebro.broker.getvalue())

if I substitute datas for data in
```
trending = bt.indicators.ADX(self.datas[0], period=14)>25
directionup = bt.indicators.PlusDI(self.datas[0], period=14) - bt.indicators.MinusDI(self.datas[0], period=14)>0
directiondn = bt.indicators.PlusDI(self.datas[0], period=14) - bt.indicators.MinusDI(self.datas[0], period=14)<0

I seem to get past the 'high' issue but then get a problem that 'trending' (and then 'buy_sig' if print(trending) is commented out) is not defined

andydoc1

Sorry, first post - how can i edit after posting???

toinou222

Hi,

You can't edit a post.

It seems that the problem comes from your dataset df1 which is not related to backtrader... It has no high column.

toinou222

My bad, didn't read right...

in next() try to access buy_sig[0] as you want to use the value not the line

run-out

@andydoc1 said in Sorry, more about missing attribute 'high', and definition in init not recognised in next:

buy_sig = bt.And(trending, directionup)
sell_sig = bt.And(trending, directiondn)

Your problem is here. The buy_sig and sell_sig are local variables to init. You need to add self in front of them to make the variables available throughout the strategy including the next method.

if self.buy_sig:
    self.log('BUY CREATE, %.5f' % self.dataclose[0])
    self.order = self.buy()
elif self.sell_sig:
    self.log('SELL CREATE, %.5f' % self.dataclose[0])
    self.order = self.sell()

run-out

@run-out I hate that I cannot edit....

Add in [0] after the indicators to ge the value for that bar.

if self.buy_sig[0]:
    self.log('BUY CREATE, %.5f' % self.dataclose[0])
    self.order = self.buy()
elif self.sell_sig[0]:
    self.log('SELL CREATE, %.5f' % self.dataclose[0])
    self.order = self.sell()

andydoc1

@toinou222 so i tried that, seemed like a possibility although the documentation says:

"A complete example which generates a buy signal during init:

class MyStrategy(bt.Strategy):

    def __init__(self):

        sma1 = btind.SimpleMovingAverage(self.data)
        ema1 = btind.ExponentialMovingAverage()

        close_over_sma = self.data.close > sma1
        close_over_ema = self.data.close > ema1
        sma_ema_diff = sma1 - ema1

        buy_sig = bt.And(close_over_sma, close_over_ema, sma_ema_diff > 0)

    def next(self):

        if buy_sig:
            self.buy()

"
with no mention of index 0 based notation

So my new 'next' is:

    def next(self):
        # Simply log the closing price of the series from the reference
        self.log('Close, %.5f' % self.dataclose[0])
        #print(self.position) 

        if buy_sig[0]:
          self.log('BUY CREATE, %.5f' % self.dataclose[0])
          self.order = self.buy()
        elif sell_sig[0]:
          self.log('SELL CREATE, %.5f' % self.dataclose[0])
          self.order = self.sell()

However I still get:

*<ipython-input-5-49c7ea4f6a06> in next(self)
39 #print(trending)
40
---> 41 if buy_sig[0]:
42 self.log('BUY CREATE, %.5f' % self.dataclose[0])
43 self.order = self.buy()

NameError: name 'buy_sig' is not defined*

run-out

Try it like this. It will margin out a lot, but basically this works.

 def __init__(self):
        sma1 = btind.SimpleMovingAverage(self.data)
        ema1 = btind.ExponentialMovingAverage(self.data)

        close_over_sma = self.data.close > sma1
        close_over_ema = self.data.close > ema1
        sma_ema_diff = sma1 - ema1

        self.buy_sig = bt.And(close_over_sma, close_over_ema, sma_ema_diff > 0)

    def next(self):
        if self.buy_sig[0]:
            self.buy()

andydoc1

@run-out so yes it does work thank you. Now I have a further problem - closing the position.

In next, I call

if self.close_sig[0]:
          self.log('CLOSE ALL, %.5f, %d' % (self.dataclose[0], self.position.size))
          self.order = self.close()  
          print(self.position)

logging self.position.size is to show me that that is an int. However, in init where I try to use it as part of my definition of self.close_sig[0] it won't allow it:

self.close_sig = bt.And(detrending
                                #, self.position.size!=0
                                )

works, but uncommenting the reference to size leads to no logging of 'CLOSE ALL' in the output (there were numerous with it commented).

The other issue I have is with the detrending definition. The following works:

detrending = bt.Or(
            bt.And(
                bt.indicators.ADX(self.datas[0]
                                  #, period=14
                                  )<=25
                   , bt.indicators.ADX(self.datas[0]
                                       #, period=14
                                       )<=bt.indicators.ADX(self.datas[-1]
                                                            #, period=14
                                                            )
                                       )
            # ,bt.And(
            #     bt.indicators.ADX(self.datas[0]
            #                            #, period=14
            #                            )<=bt.indicators.ADX(self.datas[-1]
            #                                                 #, period=14
            #                                                 )
            #        , bt.indicators.ADX(self.datas[-1]
            #                            #, period=14
            #                            )<=bt.indicators.ADX(self.datas[-2]
            #                                                 #, period=14
            #                                                 )
            #                            )
        )

but uncommenting the reference to the [-1] comparison to [-2] leads to the following error:

**---------------------------------------------------------------------------
IndexError Traceback (most recent call last)
<ipython-input-11-3bb162ac7882> in <module>()
13 print('Starting Portfolio Value: %.2f' % cerebro.broker.getvalue())
14
---> 15 cerebro.run()
16
17 print('Final Portfolio Value: %.2f' % cerebro.broker.getvalue())

4 frames
<ipython-input-10-ce52eeb03e9e> in init(self)
57 , bt.indicators.ADX(self.datas[-1]
58 #, period=14
---> 59 )<=bt.indicators.ADX(self.datas[-2]
60 #, period=14
61 )

IndexError: list index out of range**

i do not see what is wrong in either case...

run-out

@andydoc1 You should find most of your answers in the docs. Try the following:

ab_trader

@andydoc1 said in Sorry, more about missing attribute 'high', and definition in init not recognised in next:

However, in init where I try to use it as part of my definition of self.close_sig[0] it won't allow it:
self.close_sig = bt.And(detrending
#, self.position.size!=0
)

works, but uncommenting the reference to size leads to no logging of 'CLOSE ALL' in the output (there were numerous with it commented).

self.position.size is equal to zero in the __init()__. it can't be different since this is initialization of the strategy.

Read about strategy steps here https://www.backtrader.com/docu/strategy/

ab_trader

@andydoc1 said in Sorry, more about missing attribute 'high', and definition in init not recognised in next:

but uncommenting the reference to the [-1] comparison to [-2] leads to the following error:

self.datas[0] - data feed # 1 from all N data feeds added
self.datas[-1] - data feed # N from all N data feeds added
self.datas[-2] - data feed # N-1 from all N data feeds added

This is how python list indexing works.

if you use only one data feed than evidently self.datas[-2] doesn't exist. Also if single data feed is added than self.datas[0] == self.datas[-1].

I would recommend you to start reading the docs. It will save your time and help to understand the bt.