gh-149244: Support iterator inputs in covariance, correlation, and linear_regression

[htjworld]

Closes #149244.

statistics.covariance(), statistics.correlation(), and
statistics.linear_regression() previously called len() directly on their
inputs, raising TypeError for iterators and generators. This was inconsistent
with the rest of the module — mean(), variance(), and stdev() all accept
any iterable.

The fix adds x = list(x) and y = list(y) at the start of each function.
This is in line with the internal _ss() helper, whose docstring states
"Calculations are done in a single pass, allowing the input to be an iterator."
The list conversion also correctly handles repeated iteration:
covariance() iterates each input twice (fsum() then sumprod()), and
linear_regression() already notes in a comment that x must be a list
"because used three times below."

Tests for iter() and generator expression inputs are added to
TestCorrelationAndCovariance and TestLinearRegression. The documentation
for all three functions is updated to reflect that sequences or iterables are
accepted.

• Issue: statistics: covariance, correlation, and linear_regression do not accept iterator inputs #149244

[python-cla-bot]

All commit authors signed the Contributor License Agreement.

[read-the-docs-community]

Documentation build overview

📚 cpython-previews | 🛠️ Build #32501469 | 📁 Comparing e48aa88 against main (9668d26)

  🔍 Preview build  

3 files changed

± download.html
± library/statistics.html
± whatsnew/changelog.html

[htjworld]

Hi @rhettinger, could you share your thinking on closing this? I'd genuinely like to understand — whether it's the approach, a design decision I missed, or something else entirely.

[rhettinger]

It is fine for these functions to just support sequences.

Also, running list(it) just slows down the common cases with an unnecessary copy.