docs: rework the @include TIP to present both fragment-composition styles

[mpyw]

The TIP added in #23 was too strict: it claimed the including template must never own the connecting and, arguing from the un-expanded template's raw-paste validity. But an @include-containing template is an authoring artifact — Expand produces the two-way SQL — so requiring the un-expanded form to run raw was not a real constraint.

This replaces it with a TIP that presents both valid styles:

• (a) the fragment owns the leading and, and the un-expanded template is itself runnable (where 1 = 1 /*%! @include */);
• (b) the fragment is self-contained (its own 1 = 1 anchor) and the including template owns the and (where 1 = 1 and /*%! @include */), valid after Expand.

Both build to valid SQL; the author picks based on whether the un-expanded template should also run verbatim.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.52%. Comparing base (f47b2b9) to head (22e4571).

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #24   +/-   ##
=======================================
  Coverage   96.52%   96.52%           
=======================================
  Files          11       11           
  Lines         951      951           
=======================================
  Hits          918      918           
  Misses         21       21           
  Partials       12       12           

☔ 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.