Load one column of a CSV as a time series and register it under a series_id.

Register a series from inline values (small data; prefer load_csv for files).

Load a bundled synthetic sample (seeded, reproducible) — useful for demos and evals.

List every series currently loaded, with basic stats.

Distributional summary: quartiles, spread, skewness, kurtosis, missing count.

Fetch raw observations in a time window (evenly thinned if over the limit).

Resample onto a regular grid; registers and returns a NEW derived series.

Rolling-window statistics with an evenly spaced preview per stat.

Audit sampling gaps, duplicate timestamps, missing values, and regularity.

Flag anomalous observations; returns scored anomalies, strongest first.

Detect level shifts (mean changes) via CUSUM binary segmentation.

Split the series into trend/seasonal/residual and quantify each component's strength.

Run ADF and KPSS together and give a combined stationarity verdict.

ACF/PACF with significance bounds; suggests a seasonal period when one stands out.

Estimate trend three ways: OLS, robust Theil-Sen, and the Mann-Kendall test.

Correlate two series on shared timestamps and find the lag of strongest coupling.

Baseline forecast with 95% intervals and an honest holdout backtest.