Skip to content

rendering-backends.md

Latest commit

 

History

History
91 lines (70 loc) · 2.78 KB

rendering-backends.md

File metadata and controls

91 lines (70 loc) · 2.78 KB
Raw

Rendering Backends

pyoncoplot has two rendering backends:

  • Plotly for interactive output.
  • Matplotlib for deterministic static output.

Plotly

Use Plotly when you want hover labels, interactive inspection, and HTML export.

result = oncoplot(
    mutations,
    gene_col="gene",
    sample_col="sample",
    mutation_type_col="mutation_type",
    tooltip_col="tooltip",
    backend="plotly",
)

result.save("oncoplot.html")

Plotly output supports copy_on_click:

Value Copied value
"sample" clicked sample identifier
"gene" clicked gene identifier
"tooltip" clicked tooltip text
"mutation_type" clicked mutation type
"nothing" disables clipboard behavior

Clipboard behavior is inserted into exported HTML via navigator.clipboard.writeText.

Matplotlib

Use Matplotlib when you need PNG, SVG, PDF, deterministic gallery images, or static figures for manuscripts.

result = oncoplot(
    mutations,
    gene_col="gene",
    sample_col="sample",
    mutation_type_col="mutation_type",
    backend="matplotlib",
)

result.save("oncoplot.svg")

Matplotlib currently has the richest static layout support for:

  • stacked TMB bars.
  • right-side gene bars with percentage labels.
  • numeric metadata bars.
  • separate metadata legends.
  • compact tile borders and row separators.

Choosing A Backend

Need Recommended backend
exploratory browser plot Plotly
hover and click-to-copy Plotly
standalone HTML Plotly
PNG/SVG/PDF figure Matplotlib
deterministic gallery recreation Matplotlib
manuscript-style static output Matplotlib

Backend Support Matrix

Feature Plotly Matplotlib
mutation grid interactive markers with hover and linked selection static tile grid
TMB totals bar trace; log and scientific tick formatting supported bar axis; log labels and scientific tick formatting supported
TMB subtype legends TMB: <subtype> entries when custom stacked TMB is rendered separate TMB Type legend when custom stacked TMB is rendered
metadata display heatmap strip with prettified hover labels and shared interactive legend heatmap or mini-bar strip with separate metadata legends
image export HTML built in; image export needs the export extra PNG, SVG, PDF, and other savefig formats
copy-on-click supported in exported HTML not applicable to static figures
static layout controls useful font, legend, marker, and axis controls richest support for tile dimensions, metadata bars, font faces, and separate legend layout

The interactive argument is a convenience alias:

oncoplot(..., interactive=True)   # backend="plotly"
oncoplot(..., interactive=False)  # backend="matplotlib"