docs: add ClarkLDF doctest examples

[EKtheSage]

Summary: Add focused ClarkLDF Sphinx doctest examples and make the numeric comparison robust to small platform/dependency drift by rounding the displayed LDF values. Split from the larger #786 work and intentionally excludes .github/workflows/sync-main-to-docs.yml. Refs #704

---

Note

Low Risk
Documentation-only change to the class docstring; no runtime or API behavior changes.

Overview
Adds a new Examples section to the ClarkLDF class docstring in clark.py, with three Sphinx doctest blocks and narrative aimed at actuaries.

The first example fits loglogistic vs weibull on ukmotor, prints rounded LDFs and G_(240) to show tail sensitivity, and explains curve choice. The second shows Cape Cod via sample_weight and documents method_ / elr_. The third uses groupby="LOB" on the clrd sample to pool thin triangles and print omega_ / theta_.

There are no changes to estimator logic or public API—documentation and doctest coverage only.

^{Reviewed by Cursor Bugbot for commit 0377b04. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.79%. Comparing base (b24f209) to head (0377b04).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #802   +/-   ##
=======================================
  Coverage   88.79%   88.79%           
=======================================
  Files          89       89           
  Lines        5060     5060           
  Branches      646      646           
=======================================
  Hits         4493     4493           
  Misses        423      423           
  Partials      144      144           

Flag	Coverage Δ
unittests	88.79% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[henrydingliu]

i went back and read a few pages from the pandas doc. a lot of them starts with a description of what a user is looking to accomplish, then goes through the details. but if we take the most curt example docstring I found on pandas, which is the xs function, the basic flow is still "user goal" -> "code implementation", rather than "in depth description of a parameter" -> "code implementation". could we do something similar with our examples? so instead of explaining the parameter, describe an issue then offer the parameter as the solution.

[henrydingliu]

awkward phrasing. the intentionality of the user is to use cape cod. to use cape cod, the user can pass in a weight vector without additional parameters in the clarkldf transformer.

[henrydingliu]

the example text makes a claim about individual triangles being too sparse to fit. the actual code doesn't really make that apparent, or even show the results at all

[henrydingliu]

too close to be a worthy choice of an example

[kennethshsu]

@EKtheSage are you interested in finishing up this PR?

[EKtheSage]

@henrydingliu all three comments are addressed. The curve comparison now prints full ldf_ patterns plus the percent of ultimate emerged at 240 months (85% loglogistic versus 99% weibull), so the difference is material. The Cape Cod example now simply passes a premium vector as sample_weight on clrd data and prints the fitted ELRs by line. The groupby example shows the 775 company triangles being pooled into one curve fit per LOB with the fitted parameters. Main is merged in as well.