About demo algorithm 12044

I was working with this example last night so thanks to Varad Kabade for sharing work.

1. I added the rocd indiciator to the SymbolData class and the MyCoarseFilterFunction. 
    

2. Example in backtest uses heapq.heappush and heapq.nlargest to iteratively add and then more quickly search for n-largest - https://docs.python.org/3/library/heapq.html
 

import heapq
h = []

# push iteratively in each loop
heapq.heappush(h, (0.2, 'SPY'))
heapq.heappush(h, (0.1, 'TSLA'))
heapq.heappush(h, (0.3, 'AMZN'))
heapq.heappush(h, (0.7, 'APPL'))

#heapq.heappop(h)
# Grab n largest values e.g. 3 
heapq.nlargest(3,h)

Hi Ben Kaylor,
In the above algorithm, the self.rocd indicator is getting updated twice. When using:
 

self.rocd = IndicatorExtensions.Of(Delay(21), self.roc)

to update self.rocd we just need to update the self.roc. In the MyCoarseFilterFunction method, we don't need to sort the coarse objects by dollar volume.Refer to the attached backtest for the changes mentioned above.
Best,
Varad Kabade

Ben Kaylor,

It's always a pleasure for me to try something new to accomplish a task.

Thanks for the new approach.