Something like this should work for days since last high.

class Days_Since_High(CustomFactor):  
    # Define inputs  
    inputs = [USEquityPricing.high]  


    # Set window_length to whatever number of days to lookback as a default  
    # This can also be set/over-ridden as shown below:  
    # my_high_10 = Days_Since_High(window_length = 120)  
    window_length = 252  


    def compute(self, today, assets, out, high):  
        # Replace any NaNs with -inf (a low number). Sometimes the high data can be all NaNs  
        # which doesn't work even if using the nanargmax method.  
        high[np.isnan(high)] = -np.inf


        # We want to use the numpy 'argmax' method to find the indexes of the max values.  
        # However, this method only finds the first occurance. We want the last.  
        # So, flip the high prices so it's in reverse order  
        high_prices_latest_to_earliest = np.flipud(high)


        # Now simply use the 'argmax' method to find the indexes of the highs  
        # 0 will be yesterday, 1 the day before, and so on.  
        out[:] = np.argmax(high_prices_latest_to_earliest, axis=0)

The lows can be found like this

class Days_Since_Low(CustomFactor):  
    inputs = [USEquityPricing.low]  
    window_length = 252  


    def compute(self, today, assets, out, low):  
        low[np.isnan(low)] = np.inf  
        low_prices_latest_to_earliest = np.flipud(low)  
        out[:] = np.argmin(low_prices_latest_to_earliest, axis=0)  

If one wants the highest/lowest close then use 'USEquityPricing.close' instead. See attached notebook. Good luck.

Use np.argmax()

class days_since_High252(CustomFactor):  
    window_length = 252 # ~52 weeks  
    inputs = [USEquityPricing.close]  
    def compute(self, today, asset_ids, out, close_prices):  
        out[:] = 252 - np.argmax(close_prices, axis=0)  

Great, thanks so much Dan and Michael.

I just learned about numpy's argmax and argmin (the index of the max and min respectively) from one of the videos shared in a separate thread.

I really appreciate your help as I slowly start to wrap my head around numpy and the CustomFactor.