docs(gallery): derive the gallery from the notebook sources (#746)

The Gallery page duplicated the per-notebook grid cards that also live in the
notebooks submodule's section index pages (tutorials/index, examples/index).
The two lists were maintained independently in separate repos and drifted:
adding a notebook required editing both, and when one was missed the Gallery
silently omitted it (e.g. 'Interactive region annotation' on stable).

Make the section index pages the single source: gallery.md keeps the intro and
a visible toctree to those pages, so every notebook is listed automatically and
the Gallery can never fall out of sync. Remove the now-orphaned card images
under _static/gallery/ (the section cards use _static/img/).

Fixes #746.

Author: timtreis

docs/_static/gallery/color_and_palette.png

@@ -8,73 +8,14 @@ Sources live in
 every notebook is executable end-to-end and re-executed on a weekly schedule
 against the latest `spatialdata-plot` release.
 
-## Tutorials
+- **Tutorials** — learn the API on synthetic data. Start with *Getting started*.
+- **Examples** — worked analyses on datasets you'd actually encounter.
 
-Entry-point material for learning the API on synthetic data.
-
-::::{grid} 1 2 2 2
-:gutter: 3
-
-:::{grid-item-card} Getting started
-:link: notebooks/tutorials/getting_started
-:link-type: doc
-:img-top: _static/gallery/getting_started.png
-
-The fluent `.pl` API, layering, and styling on the in-memory `blobs`
-dataset. Ideal first read.
-:::
-
-:::{grid-item-card} Colour and palettes
-:link: notebooks/tutorials/color_and_palette
-:link-type: doc
-:img-top: _static/gallery/color_and_palette.png
-
-How `color=` resolves, the v0.3.0 `groups` behaviour, and building
-perceptually well-spaced or colourblind-safe palettes with
-`make_palette` and `make_palette_from_data`.
-:::
-
-:::{grid-item-card} Speeding up rendering
-:link: notebooks/tutorials/performance
-:link-type: doc
-:img-top: _static/gallery/performance.png
-
-Keep rendering fast on large data: automatic rasterization and scale
-selection for images, and the `datashader` backend for large collections
-of shapes and points.
-:::
-
-::::
-
-## Examples
-
-Worked examples on real datasets you'd actually analyse.
-
-::::{grid} 1 2 2 2
-:gutter: 3
-
-:::{grid-item-card} Visium mouse brain
-:link: notebooks/examples/visium_mouse_brain
-:link-type: doc
-:img-top: _static/gallery/visium_mouse_brain.png
-
-Render H&E tissue, overlay spots, color by gene expression and by cluster,
-and finish with a publication-style figure.
-:::
-
-:::{grid-item-card} Interactive region annotation
-:link: notebooks/examples/interactive_annotate
-:link-type: doc
-:img-top: _static/gallery/interactive_annotate.png
-
-Draw regions of interest directly on a `spatialdata-plot` canvas with
-`sdata.pl.annotate(...)` and persist them as a `ShapesModel` element.
-:::
-
-::::
+Each section below links every notebook it contains. The card lists live with
+the notebook sources (the section index pages), so a newly added notebook shows
+up here automatically and the gallery can never fall out of sync.
 
 ```{toctree}
-:hidden:
 :maxdepth: 2
 
 notebooks/tutorials/index