docs: clarify that H.edges[idx] returns attributes, not members

[leotrs]

Summary

A recurring point of confusion is that H.edges[0] returns the (empty) attribute dict rather than the members of edge 0. The behavior is technically correct, but the docstring didn't direct users to the actual member-access methods.

This PR keeps the behavior the same (per our v1.0 stance of hardening rather than changing the API) and improves the docstring:

• Upfront note saying explicitly that this returns attributes, not members
• Concrete examples showing both the empty default and a case with attributes set
• See Also pointing at EdgeView.members and NodeView.memberships

Test plan

• Doctests pass (run with --doctest-modules)
• Full test suite passes (415 passed, 6 skipped)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.68%. Comparing base (b7a899d) to head (2c92ffb).
⚠️ Report is 4 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #727   +/-   ##
=======================================
  Coverage   93.68%   93.68%           
=======================================
  Files          66       66           
  Lines        5207     5213    +6     
=======================================
+ Hits         4878     4884    +6     
  Misses        329      329           

☔ View full report in Codecov by Sentry.
📢 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.

[maximelucas]

Great idea to clarify this, thanks Leo.
I wonder though in what context the user will actually see the docs of the .getitem() method?
Like do we get there by typing H.nodes and then in notebook shift+tab to see the docs, or ..?
This is just to make sure that users will see this explanation easily and often.

[leotrs]

Good point Max. Strictly speaking the __getitem__ docstring only shows up via H.edges.__getitem__? or on the API reference page, not via the more natural H.edges[0]? (which evaluates first and shows the dict's docstring instead).

Pushed a follow-up that also adds a short attr-vs-member note to the NodeView and EdgeView class-level docstrings. Now H.nodes? / H.edges? (a common discovery move) surfaces the distinction too.

[maximelucas]

Great!

[maximelucas]

Oops, this was merged into main rather than dev. @leotrs @kaiser-dan ?
Btw, by default a new PR merges into main, can we change a setting so that the default is into dev?

[leotrs]

Sorry about that, totally my mistake — I missed the dev/main convention. Just retargeted #728 and #729 to dev, backmerged main into dev to recover the stranded commits, and opened #734 to document the workflow in CLAUDE.md so it doesn't happen again. Going to change the default branch to dev now so PRs target dev by default.