name: visualization
description: "Render Squidpy plots for spatial overlays, segmentation, graph
statistics, ligand-receptor heatmaps, and variation-by-distance results."
disable-model-invocation: true
metadata:
disco-role: operating
license: BSD 3-Clause
Squidpy Visualization
Use this sub-skill when a task is about rendering Squidpy figures after the AnnData or SpatialData-like table is already loaded and, when needed, analysis results have already been computed.
Route
- Spatial overlays and segmentation masks: read
references/plotting-workflows.md#spatial-overlays and references/api-reference.md#spatial-plotters.
- Neighborhood, centrality, interaction, co-occurrence, Ripley, or ligand-receptor plots: read
references/plotting-workflows.md#graph-and-statistic-plots.
- Distance-to-anchor plots and image-feature extraction for plotting: read
references/plotting-workflows.md#tool-and-feature-plotting.
- Plotting failures, headless rendering, palette mismatches, saved files, or visual-regression differences: read
references/troubleshooting.md.
- Smoke-test a plotting installation without downloads or repository fixtures: run
python scripts/plotting_smoke.py --quiet.
Boundaries
- This sub-skill renders figures with
squidpy.pl; it does not compute graph statistics, ligand-receptor tests, or distance design matrices.
- Route graph construction and statistic computation to
../graph-analysis/SKILL.md before plotting graph outputs.
- Route design-matrix creation for
sq.pl.var_by_distance to ../tools-workflows/SKILL.md.
- Route data loading, Visium metadata repair, missing
.obsm['spatial'], or missing .uns['spatial'] image metadata to ../datasets-and-io/SKILL.md.
- Route stable
ImageContainer processing to ../image-analysis/SKILL.md and experimental QC or tiling plots to ../experimental-imaging/SKILL.md.
Practical Defaults
- Use
matplotlib.use("Agg") before importing pyplot in headless scripts.
- Prefer
return_ax=True for sq.pl.spatial_scatter, sq.pl.spatial_segment, and sq.pl.var_by_distance when the caller needs axes.
- For graph/statistic plotters, compute the corresponding
sq.gr.* result first and treat plot functions as renderers over stored keys.
- For image overlays, verify
library_id, library_key, img_res_key, scalefactors, and segmentation cell IDs before tuning aesthetics.