How to dynamically inject logic in a strategy at runtime?

[binarymason]

Hi there!

Thanks for the wonderful library. I've been experimented with it for a while now, and it has been great.

Right now I'm trying to integrate an existing code base that has a concept of signals, and in order to avoid duplicated code logic, I'm trying to work with python inheritance or class factories to achieve this.

Unfortunately, when I run the backtest no trades get executed. Any insights would be appreciated!

---

Here's an example Signal that I'm trying to make work with a backtesting.Strategy:

class Signal():
  def __init__(self, symbol, data, **params):
    self.symbol = symbol
    self.data = data
    self.params = dict(params, **self.defaults())
    self.setup()
    
  def __getattr__(self, name):
    """Look in params if attribute is missing"""
    attr = self.__dict__.get("params", {}).get(name)
    if attr is not None:
      return attr
    else:
      raise AttributeError(name)

  @abstractmethod
  def setup(self):
    pass
  
  @abstractmethod
  def buy(self):
    pass

  @abstractmethod 
  def sell(self):
    pass
    
  def I(self, func, *args, **kwargs):
    """attempt to have same signature as backtesting.py"""
    return func(*args, **kwargs)
  
  def defaults(self):
    """returns first value in param combos"""
    store = {}
    for k, v in self.param_combos().items():
      store[k] = v[0]
    return store
    
  def param_combos(self):
    return {}
  
  
class SmaCross(Signal):
  def setup(self):
    self.sma1 = self.I(SMA, self.data.Close, self.n1)
    self.sma2 = self.I(SMA, self.data.Close, self.n2)

  def next(self):
    # If sma1 crosses above sma2, close any existing
    # short trades, and buy the asset
    if crossover(self.sma1, self.sma2):
      self.position.close()
      self.buy()

    # Else, if sma1 crosses below sma2, close any existing
    # long trades, and sell the asset
    elif crossover(self.sma2, self.sma1):
      self.position.close()
      self.sell()
      
      
  def param_combos(self):
    return dict(n1=range(5, 30, 5), n2=range(10, 70, 5))
    

Then to create a backtesting.Strategy, I am currently running this factory code:

from backtesting import Strategy
from signals import SmaCross
import pandas as pd

df = pd.read_csv("data.csv")

sig = SmaCross("AAPL", df)

MyBacktest = type('MyBacktest', (Strategy,), dict(
  sig.params, init=sig.setup, next=sig.next
))

bt = Backtest(df, MyBacktest, cash=100_000, commission=.02)
stats = bt.run()
print(stats['# Trades'])

# >>> 0

If I copy and paste the demo from online and use it on the same dataset with same parameters, there are 34 trades, so I know there's not a problem with my dataset or parameter values. Something is wrong with the custom type class.

Thanks for your help!

[kernc]

MyBacktest = type('MyBacktest', (Strategy,), dict(
  sig.params, init=sig.setup, next=sig.next
))

You're not calling super() methods in your overriding "init" and next. Strategy methods never fire.

[binarymason]

Thanks kernc!