Module summary 

describe / summarize logic (Phase 21).

Free functions over a &rusqlite::Connection, called by SqliteBackend::summarize. The numeric statistics reproduce pandas .describe() exactly so a notebook user can cross-check the agent’s output against their own dataframe (docs/design/centaur-data-scientist.md): the sample standard deviation (ddof = 1) and linear-interpolation quartiles (the numpy/pandas default).

Functions

describe

pandas .describe() over a sorted, non-null f64 slice (Phase 21).

quantile_sorted

The q-quantile of a sorted slice by linear interpolation — the numpy/pandas default (interpolation="linear"): pos = (n - 1) * q, lower = floor(pos), frac = pos - lower, val = v[lower] + frac * (v[lower + 1] - v[lower]).

summarize

Summarize table: schema, null/distinct counts, and pandas-style describe for numeric columns. DataError::NoSuchTable if it does not exist.