Module stats 

Statistics: streaming (Welford) and batch algorithms.

Modules

chart_overlay 🔒

Single-call computation of all chart statistical overlays for one WASM boundary crossing.

simd

SIMD-accelerated helpers for telemetry hot paths (WASM SIMD128 on wasm32+simd128, autovectorized scalar elsewhere).

summary 🔒

Statistics: streaming (Welford) and batch algorithms.

timeline_layout 🔒

Timeline layout helpers shared by time-vs-event UI (tick placement and lane packing).

Structs

Batch

Batch statistics for a complete dataset.

ChartOverlayView

Pre-computed chart overlay data returned across the WASM boundary.

LinearRegression

ScatterOverlayView

Pre-computed scatter overlay (mean lines + regression segment); when valid is false the numeric fields are zeros.

Streaming

Streaming mean/variance via Welford’s online algorithm (O(1) memory).

TimelineGeometry

Layout + viewport inputs for a timeline (pixels and milliseconds; zoom/pan_ms are clamped by the math).

TimelineMetrics

Derived pixel/time geometry for one render of a timeline.

TimelineViewport

A timeline viewport: the committed pan offset and zoom factor.

Constants

DEFAULT_TARGET_PIXELS_PER_TICK

Target pixel gap between adjacent ticks when no caller preference is set.

Functions

assign_timeline_lanes

Pack (start, end) intervals into non-overlapping lanes via greedy scheduling, returning lanes in input order.

compute_all_percentiles_from_hdr

Compute all standard percentiles from an HDR histogram.

compute_bucket_averages_x10

Compute per-bucket dps_per_bucket_x10 averages (sum/samples, 0 when samples == 0).

compute_chart_overlay

Compute all chart overlay statistics in one call.

compute_result_core

Compute (mean_dps_x10, std_dps_x10, ci95_half_pct_x10000) for ResultCoreV1.

compute_scatter_overlay

Compute scatter overlay statistics from paired (x, y) values; non-finite pairs are dropped, valid: false if too few pairs remain or xs are identical.

compute_timeline_metrics

Derive all pixel/time geometry (including axis ticks) for one timeline render; the source of truth for px<->ms.

compute_timeline_ticks

Pick “nice” tick positions (ms) inside [0, duration_ms] for a timeline axis.

correlation

Compute Pearson correlation coefficient.

covariance

Compute sample covariance between two slices.

ema

Exponential moving average with smoothing factor alpha (0-1).

ema_span

EMA with span-based smoothing: alpha = 2 / (span + 1).

histogram_percentile_from_hdr

Compute a percentile (p in 0.0..=1.0) from an HDR histogram, in x10 space.

linear_regression

Compute simple linear regression using least squares.

merge_dps_stats

Merge chunk DPS stats into the running aggregate via Chan/Golub/LeVeque pairwise merge.

pan_timeline_by

Compute the new viewport after dragging the track by delta_px pixels (zoom unchanged).

sma

Simple moving average with window size.

zoom_timeline_at

Compute the new viewport after a cursor-anchored zoom step (local_x pixels from the track left, cursor time fixed).

Type Aliases

Summary

Alias for Batch.