Per-panel y-limits (and clearer share_y) for collected summary facet plots

[jepegit]

Problem / context

Collected summary plots facet variables (capacity, CE, …) with _cycles_plotter(match_axes=True) by default, so all rows share one y-scale. CE outliers (~1e6) crush other panels. Downstream apps (e.g. cellpy-simple-gui) need independent auto-scale and per-panel fixed limits.

match_axes=False already unmatches axes. AxisSpec.range / PanelSpec.y_axis exist, and some render paths have y_range / ce_range, but there is no clear, documented API to set per-facet-row y-limits on the collected summary path and have the Plotly backend apply them reliably.

Companion (cellpy-simple-gui): cellpy/cellpy-simple-gui#2

Spec

• Document and stabilize match_axes / share_y for summary facets (independent vs shared).
• Support per-panel (per-variable) y-range on collected summary figures end-to-end (spec → backend), not only a single global y_range.
• Prefer keeping auto-range per panel when limits are omitted.

Acceptance criteria

• Public API (kwargs or FigureSpec/PanelSpec) can set independent y-scales for summary facets.
• Public API can set per-panel y-limits; Plotly (and ideally MPL) honor them.
• Tests cover shared vs independent and per-panel range application.
• Docs/examples show Capacity+CE with a CE limit or independent scales.

Context

Requested while improving cellpy-simple-gui summary plotting.

[jepegit]

Companion app issue (cellpy-simple-gui): cellpy/cellpy-simple-gui#2