Clarify package architecture: marks in netrics, is_*() in manynet, dependency guidance for autograph

[Copilot]

Addresses the question of where mark functions should live and how downstream packages like {autograph} should structure their dependencies.

Decision: node_is_*() and tie_is_*() stay in {netrics} — they are analytic functions (finding cutpoints, bridges requires graph algorithms). Network-level is_*() tests (is_directed(), is_twomode()) stay in {manynet} — they describe data type, not analytic results. {autograph} should depend on both since it needs network classes and analytic results to visualize.

Changes

• README.Rmd/README.md: Rewrite About and Relationship sections to define the stocnet dependency graph (manynet → netrics → autograph/migraph) and the data-property vs analytic-result separation principle
• README.Rmd/README.md: Separate Measures (*_by_*()) from Motifs (*_x_*()) into distinct sections — these were conflated under a single "Motifs" heading
• README.Rmd/README.md: Fix stale function names in Measuring section (node_degree() → node_by_degree(), etc.)
• DESCRIPTION: Enumerate the four function families with naming patterns in package description
• _pkgdown.yml: Add naming conventions to section descriptions (Marking → node_is_*(), Measures → *_by_*(), etc.)

Architecture

{manynet}    → classes, coercion, modification, is_*()
{netrics}    → measures _by_(), memberships _in_(), motifs _x_(), marks node_is_/tie_is_()
{autograph}  → visualization — depends on both manynet + netrics
{migraph}    → modelling — depends on both manynet + netrics

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.