Struct TimedWindowFilter

pub struct TimedWindowFilter<T, K: KernelGenerator> { /* private fields */ }

Available on crate feature std only.

Time-series data downsampling based on a convolution kernel. Stores up to 2x the time window for nyquist sampling reasons. Once it has a full time window duration, will run the kernel and provide the result of the convolution operation.

Implementations

impl<T: Copy, K: KernelGenerator> TimedWindowFilter<T, K>

pub fn new( window_duration: Duration, bin_strategy: WindowBinStrategy, kernel_generator: Box<dyn KernelBuilder<Output = K>>, ) -> Self

pub fn add_sample(&mut self, sample: Sample<T>) -> Option<Sample<T>>

pub fn insert(&mut self, time: Timestamp<T>, value: f64) -> Option<Sample<T>>

Push a new sample into the filter. If there’s sufficient data to run the downsampling, will run and return the result.