🤖 Repo Assist — automated AI improvement.

Summary

Adds Series.pctChange and Frame.pctChange functions that compute the relative (percentage) change between each value and the value at a given offset:

result[k] = (series[k] - series[k - offset]) / series[k - offset]

This is the standard way to compute financial returns (e.g. daily stock returns from a price series) and is the F# equivalent of pandas' pct_change.

Motivation

The homepage example (see #517) attempts to calculate daily returns using Frame.diff 1 / df, which is incorrect because diff gives the absolute change and dividing by df gives something subtly different from (df - shift(df,1)) / shift(df,1). A dedicated pctChange function removes this confusion and simplifies the idiomatic usage.

Before (current — from homepage docs):

let past = df |> Frame.diff 1
let rets = past / df * 100.0   // Wrong: divides diff by current, not by previous

After:

let rets = df |> Frame.pctChange 1   // Correct: (curr - prev) / prev

Changes

Test Status

✅ All 498 existing tests pass
✅ All 5 new pctChange tests pass
Build: no new warnings introduced by these changes

Generated by Repo Assist · ◷

To install this agentic workflow, run

gh aw add githubnext/agentics/workflows/repo-assist.md@30f2254f2a7a944da1224df45d181a3f8faefd0d