API Reference - Algorithms

WSDP provides a comprehensive algorithm library for CSI processing with a pluggable architecture.

Quick Reference

Category	Built-in Algorithms
Denoising	wavelet, butterworth, savgol
Calibration	linear, polynomial, stc, robust
Normalization	z-score, min-max
Interpolation	linear, cubic, nearest
Features	doppler, entropy, ratio, decomposition
Detection	activity, change_point

Unified API

from wsdp.algorithms import denoise, calibrate, normalize, interpolate, extract_features

denoised = denoise(csi, method='butterworth', order=5)
calibrated = calibrate(denoised, method='stc')
normalized = normalize(calibrated, method='z-score')
features = extract_features(normalized, features=['doppler', 'entropy'])

Pluggable Architecture

Register Custom Algorithms

from wsdp.algorithms import register_algorithm, denoise

def my_denoise(csi, **kwargs):
    return csi * 0.5

register_algorithm('denoise', 'my_method', my_denoise)
result = denoise(csi, method='my_method')

Use Presets

from wsdp.algorithms import apply_preset, execute_pipeline

steps = apply_preset('high_quality')
processed = execute_pipeline(csi, steps)

Load from Config File

from wsdp.algorithms import load_config, execute_pipeline

config = load_config('algorithms_config.yaml')
processed = execute_pipeline(csi, config)

See the Full API Reference for complete documentation of all algorithms, parameters, and references.