deephaven.MovingAverages

Deephaven moving averages. See the java classes in com.illumon.numerics.movingaverages for more details on Deephaven moving averages.

ByEmaSimple(nullBehavior, nanBehavior, mode, timeScale, timeUnit, type=None)

Constructor for a DB aware Exponential Moving Average (EMA) which performs a by ema calculation without the added inefficiency of explicitly performing the grouping and then ungrouping operations.

Parameters

  • nullBehavior – enum value ‘BD_RESET’, ‘BD_SKIP’, ‘BD_PROCESS’ which determines calculation behavior upon encountering a null value.

  • nanBehavior – enum value ‘BD_RESET’, ‘BD_SKIP’, ‘BD_PROCESS’ which determines calculation behavior upon encountering a null value.

  • mode – enum value ‘TICK’, ‘TIME’ specifying whether to calculate the ema with respect to record count or elapsed time.

  • timeScale – the ema decay constant.

  • timeUnit – None (assumed Nanoseconds), or one of the java.util.concurrent.TimeUnit enum values - like ‘MILLISECONDS’, ‘SECONDS’, ‘MINUTES’, ‘HOURS’, …

  • type – None or enum value ‘LEVEL’, ‘DIFFERENCE’.

Returns

com.illumon.numerics.movingaverages.ByEmaSimple instance.

Ema(type, mode, timeScale)

Constructor for a Exponential Moving Average (EMA) calculation object.

Parameters

  • type – None or enum value ‘LEVEL’, ‘DIFFERENCE’.

  • mode – enum value ‘TICK’, ‘TIME’ specifying whether to calculate the ema with respect to record count or elapsed time.

  • timeScale – the ema decay constant (wrt nanosecond timestamp).

Returns

com.illumon.numerics.movingaverages.Ema instance.

EmaArray(type, mode, timeScales)

Constructor for object managing an array of Exponential Moving Average (EMA) objects.

Parameters

  • type – enum value ‘LEVEL’, ‘DIFFERENCE’.

  • mode – enum value ‘TICK’, ‘TIME’ specifying whether to calculate the ema with respect to record count or elapsed time.

  • timeScales – the ema decay constants (wrt nanosecond timestamp) - list, tuple, numpy.ndarray, or java double array.

Returns

com.illumon.numerics.movingaverages.EmaArray instance.

ExponentiallyDecayedSum(decayRate, enableTimestepOutOfOrderException=True)

Constructor for an object to calculate a sum where the values are decayed at an exponential rate to zero.

Parameters

  • decayRate – (double) rate in milliseconds to decay the sum.

  • enableTimestepOutOfOrderException – (boolean) true to allow an exception to be thrown when timesteps are not sequential.

Returns

com.illumon.numerics.movingaverages.ExponentiallyDecayedSum instance.

NeighborEmaArrayDifferencer(emaMode, emaTimescales, includeDifferenceWithCurrentValue)

Constructor for an object which manages and differences an array of emas.

Parameters

  • emaMode – enum value ‘TICK’, ‘TIME’ specifying whether to calculate the ema with respect to record count or elapsed time.

  • emaTimescales – timescales for the emas (wrt nanoseconds) - list, tuple, numpy.ndarray, or java double array.

  • includeDifferenceWithCurrentValue – true includes the difference between the latest input value and the first ema in the output results; false does not.

Returns

com.illumon.numerics.movingaverages.NeighborEmaArrayDifferencer instance.

TradingHoursEma(type, mode, timeScale, tradingHours, gapTolerance=None)

Constructor for a simple ema calculation object which decays with trading time and only accepts samples during trading hours.

..Note: This is intended to be used when sampling is frequent.

Parameters

  • type – enum value ‘LEVEL’, ‘DIFFERENCE’.

  • mode – enum value ‘TICK’, ‘TIME’ specifying whether to calculate the ema with respect to record count or elapsed time.

  • timeScale – timescale for the EMA.

  • tradingHours – object which implements (com.illumon.numerics.movingaverages.TradingHours) interface.

  • gapTolerance – double - time gap beyond which to restart the ema calculation - e.g. for a new trading day.

Returns

com.illumon.numerics.movingaverages.TradingHoursEma instance.