Add statistics nodes (median, variance, std dev, …)

[xujustinj]

Motivation

Beyond basic arithmetic (#153), workflows over numeric sequences commonly need summary statistics. We already have Sum and (proposed) Mean/Min/Max; this adds the rest. Python's stdlib statistics module covers most of these directly.

Proposed nodes

All take a single SequenceValue[FloatValue] input and output a FloatValue unless noted.

• Median — middle value; for even-length sequences, define interpolation (recommend statistics.median, i.e. mean of the two middle values).
• Mode — most common value. Define tie behavior (statistics.mode raises on no unique mode in older Pythons; multimode returns all — pick one and document).
• Variance — with a population (bool) param: sample (statistics.variance, n-1) vs population (pvariance, n). Default to one and document.
• StandardDeviation — same population param (stdev / pstdev).
• Range — max - min → FloatValue.
• Percentile / Quantile — a q param (0–100 or 0–1); define interpolation method (statistics.quantiles or numpy-style).
• Count / Length — number of elements → IntegerValue. (Overlaps with a generic sequence Length; see the sequence-utilities issue — decide whether stats reuses that or has its own.)

Conventions / decisions

• Empty sequence → raise NodeException (consistent with the decision in Add more built-in arithmetic nodes #153 for Min/Max/Mean).
• Default FloatValue in/out; Integer→Float casts already handle integer sources.
• Required title/description; auto-derived type; version 1.0.0.
• Sample-vs-population and percentile-method defaults should be explicit params, not hidden.

Related

• Companion to Add more built-in arithmetic nodes #153 (arithmetic). Not blocked by Support type-union ports (e.g. FloatValue | SequenceValue[FloatValue]) #154 — these are always sequence-input.
• Benefits from Exploration: back FloatValue with Decimal for exact decimal arithmetic #151 (Decimal) for exactness, though most stats involve division so tolerance still applies.