Module helpers 

Calculations on scalar series for built-in and host-defined studies.

Outputs retain the input length. Missing results are NaN; lengths of zero are treated as one, except change, where zero means no change. Complete-window calculations require finite inputs throughout the window. Recursive averages skip non-finite inputs, emitting NaN there and retaining their previous state.

Functions

change

Difference from period bars ago. Unavailable or non-finite endpoints yield NaN.

crossover

True where a moves from below b to above it. Equal values retain the last unequal comparison, so a touch and return does not cross. Missing values break comparisons. Output has a.len() entries; missing b entries are gaps.

crossunder

True where a moves from above b to below it; otherwise as crossover.

ema

Exponential average with alpha 2 / (period + 1), seeded by the first finite value.

ema_sma

Exponential average with alpha 2 / (period + 1), seeded by the mean of the first period finite values. Useful for nested averages with a full warm-up.

highest

Highest value in the trailing window, including partial warm-up windows. NaN never wins; an all-NaN window yields negative infinity, as f64::max does.

lowest

Lowest value in the trailing window, including partial warm-up windows. NaN never wins; an all-NaN window yields positive infinity, as f64::min does.

rma

Wilder’s average with alpha 1 / period, seeded by the mean of the first period finite values.

sma

Arithmetic mean of each complete trailing window.

stdev

Population standard deviation of each complete window, with divisor period.

sum

Sum of each complete trailing window, recovering after missing inputs leave it.

wma

Linear weighted mean of each complete window, weights 1..=period, newest heaviest.