Oleg Polakow
Oleg Polakow
@emiliobasualdo ```python vbt.IndicatorFactory(input_names=['open', 'high'], param_names=['x'], output_names=['out'])\ .from_apply_func(njit(lambda open, high, x: (open * 2 - high) < x))\ .run(pd.Series([1, 2, 3]), pd.Series([1, 2, 3]), [1, 2]).out custom_x 1 2 0 False...
I can’t access your notebook. But shapes (1080, 320) and (1080, 5) are not broadcastable in numpy (and shouldn’t be). In vectorbt, they may still broadcast if their column levels...
This worked because vectorbt does the broadcasting (https://polakowo.io/vectorbt/docs/index.html?#broadcasting and https://polakowo.io/vectorbt/docs/base/reshape_fns.html#vectorbt.base.reshape_fns.broadcast). Broadcasting based on column levels is one of the main features of vectorbt.
Hi @trevor-richardson, you can pass `is_open_safe` directly to `OHLCSTEX.run`, but note that it has no effect if `wait` is set to 1 (default), so you can pass any `open` and...
If by inverse contract you mean trading “USD/BTC” instead of “BTC/USD”, there is no change needed apart from downloading the right data, vectorbt can work with any quote and base...
Good start, native enum classes are indeed much more intuitive than named tuples, just make sure to adapt https://github.com/polakowo/vectorbt/blob/master/vectorbt/utils/enum.py, also `is_namedtuple` in https://github.com/polakowo/vectorbt/blob/master/vectorbt/utils/checks.py is currently being used for checking enums,...
@RileyMShea hmm interesting observation, I just tried to come up with some hack but was unable to find any. Your composition approach makes a lot of sense given that all...
@RileyMShea it's fine, you can jump in anytime, we're not in any hurry :) Wish you a stress-free holiday!
Look-ahead indicators are rolling min, max, etc. but in the reversed direction, such that you can use them to generate labels for machine learning. For example, give me 1 if...
Plot shows your price overlayed with a heatmap of generated labels.