Creating a new class and inherit QCAlgorithm

Hi Freidavid,

Your main algorithm must inherit from QCAlgorithm. Like this:

class MyMainAlgo(QCAlgorithm):
	pass

And the error with your Pair class is that you create a RollingWindow[TradeBar] but then you Add float in it (close values), not TradeBar.

Fred

Hey Fred

Thank you so much for your answer. Unfortunately I always get an error when inheriting the QCAlgorithm in the pairs class. I've fixed the datatype of the Window to “float” thank you

The problem was that I didn't have access to the Add() method of the window. I could solve this problem just by renaming the window. It was a problem that I used a number in the name for a self.attribute.

Thank yu so much!

You are welcome Freidavid.

One day, we will have a book on LEAN. In the meantime, I'll only say that I don't think you want your class Pair to inherit from QCAlgorithm. QCAlgorithm is the class used by the LEAN engine to basically structure EVERYTHING for you, all the logic behind the strategy/all the data you need to act upon.

Pair is just a data structure for you that you are going to use in your future algo. At least, this is my understanding.

From the package you import, you are not going to do a simple strategy 😊 so I certainly don't have the complete picture here. But try the following code for Pair. My “instinct” is telling me that is what you intend to have:

class Pair:
#holds the symbols of the 2 two chosen stocks
#calculates the spread (w/o given alpha/beta) for a time period
#calculates the spread (w/ given alpha/beta) for a time period
#holds a rolling window of the price history to update in the OnData method

    def __init__(self, s1, s2, lookback, history_s1, history_s2):
        self.pair_symbols = (s1,s2)
        self.s1_closeWindow = RollingWindow[float](lookback)
        self.s2_closeWindow = RollingWindow[float](lookback)
        
        
        df_s1 = np.log(history_s1.loc[s1]['close'])
        df_s2 = np.log(history_s2.loc[s2]['close'])
        
        for close_s1 in df_s1: # I removed the list copy (i.e. whole list slice) - don't think it is needed (?)
            self.s1_closeWindow.Add(close_s1) 
        for close_s2 in df_s2: # I removed the list copy (i.e. whole list slice) - don't think it is needed (?)
            self.s2_closeWindow.Add(close_s2)
            
        self.lenght_s1 = len(self.s1_closeWindow)
        self.lenght_s2 = len(self.s2_closeWindow)

From what I recall, you had access to the Add method of the rolling windows BUT your parameterized types did not match so you were getting a TypeError between TradeBar and decimal (we were in C# if I recall right - so TradeBar and float in Python). RollingWindow is a generic class. When you create RollingWindow[TradeBar], you cannot then do Add(decimal). It won't exist, especially in C# (strongly-typed language).

Fred

Okey, I'll fix that with the Tradebar and Float, thank you very much:)