Add targeted coverage for ergonomics helpers and profiling edge cases

[Copilot]

This PR addresses test coverage gaps in helper-layer APIs by adding focused tests for un/under-tested behavior in py3plex.ergonomics and py3plex.profiling. The changes are scoped to tests only and target correctness-sensitive control paths (dispatch, filtering/output shape, and benchmark statistics).

• Ergonomics helper coverage (tests/test_ergonomics_helpers.py)

  • Added tests for quick_network:
    • replica/node construction across layers
    • intra-layer and inter-layer edge wiring
    • directed/undirected behavior
  • Added tests for quick_analysis:
    • returned payload contract (dataframe, count, network_stats)
  • Added tests for quick_communities:
    • invalid algorithm error path
    • algorithm dispatch for both louvain and leiden via monkeypatched mocks
  • Added test for show_network_summary:
    • verifies key summary sections/fields are emitted

• Profiling edge-case coverage (tests/test_profiling_additional.py)

  • Added empty-monitor report assertion for PerformanceMonitor.get_report()
  • Added deterministic median-path tests for benchmark():
    • even-iteration median (average of middle two)
    • odd-iteration median (middle element)

• Representative snippet

  mock_louvain = Mock(return_value={("A", "layer1"): 0, ("B", "layer1"): 1})
  monkeypatch.setattr(cd, "louvain_multilayer", mock_louvain)
  
  result = quick_communities(net, algorithm="louvain", seed=123)
  mock_louvain.assert_called_once_with(net, random_state=123)
  assert result["n_communities"] == 2