Summary
RainbowOscillator is RTTA's streaming implementation of Mel Widner's rainbow
oscillator: the percent width of the recursive SMA rainbow relative to price,
plus the percent position of price inside that band.
Update API
result = rtta.RainbowOscillator(period=2, layers=10, fillna=True).update(price)
# result.value, result.position, result.width
The update(...) call consumes one price observation. advance(...) updates
state without returning a Python object. Multi-output batch(...) returns
arrays for value, position, and width.
Theory Of Operation
The oscillator reuses the same recursive SMA stack as RainbowMovingAverage.
When the rainbow is wide relative to price, trend dispersion across lag depths
is large; when it is narrow, the layers have converged. RTTA also reports where
price sits between the highest and lowest rainbow layers, which is useful as a
normalized location feature.
Recurrence
Let \(x_t\) be price and let \(H_t\), \(L_t\), \(\operatorname{width}_t = H_t-L_t\),
and \(\operatorname{mid}_t = \tfrac12(H_t+L_t)\) be the rainbow envelope from
RainbowMovingAverage with the same period and layers.
Division by zero (flat rainbow or zero price) is handled by the library's safe divide and returns \(0\).
Implementation Notes
The recurrence is implemented in src/rtta/indicator.cpp in
class RainbowOscillator, which owns an internal RainbowMovingAverage.