[FEATURE] Apply automated grid search straight from backtest

[enzoampil]

Currently the backtest function only takes on one set of arguments at a time. In practice we actually want to run results on multiple combinations. (e.g. multiple possible values for slow and fast moving average).

Utilize cerebro.optstrategy to perform the above (reference)

[jpdeleon]

Once done, we can mention in the notebook that fastquant does this automatically. :)

I think this merits backtest to be a class of its own so we access it's properties e.g. the best parameters, the backtrader figure using the best parameters, etc.

[enzoampil]

@jpdeleon Noted on the notebook :)

hmm not necessary for it to have its own class if all of these properties are found straight from the cerebro object rigth?

[jpdeleon]

@enzoampil I agree. Sorry I missed earlier your note about cerebro.optstrategy that does parameter optimization under the hood already. I think any parameter in a strategy can be optimized already using that method.

So the grid_search notebook is only useful for presentation but not useful in practice.

Also check this backtrader script that optimizes across strategies.

[enzoampil]

Planning to apply the ff:

1. Perform grid search if any of the parameters are input as an iterable (kwargs input)
2. If plot = True only the "optimal" version of the strategy is returned
3. Add explicit usage of Analyzers to choose the metric of comparison across strategies (ref)
4. Will set returns and sharpe as the default analyzers.
5. Instead of cerebro, parameters and metrics will be returned by backtest. These come from stratruns, since this contains both analyzer metrics and all metrics related to the run (e.g. final portfolio value)