Inline plots¶

PySofra can attach a matplotlib-rendered plot to a SofraTable and serialise it once into SVG + PNG + PDF, so each renderer picks the form it supports:

Renderer	Plot format used
`to_html()`	inline `<svg>` (rendered above or below the table)
`to_docx()`	embedded PNG (via `python-docx`'s `add_picture`)
`to_pptx()`	embedded PNG (placed on the slide alongside the table)
`to_latex()`	embedded PDF (written as a sidecar file referenced by `\includegraphics`)
`to_xlsx()`	not embedded (Excel users typically prefer numeric data)
`to_image()`	not embedded (the table-as-image path is itself a PNG)

All three byte streams are post-processed so that two renders of the same plot are byte-identical — matplotlib's wall-clock timestamps, process-randomised DOM IDs, and PDF creation dates are stripped so publication artifacts remain reproducible across processes and CI runs.

Forest plot for regression tables¶

ps.tbl_regression(fit, exponentiate=True).with_forest_plot()

Reads point estimates and CI bounds from the displayed cells, so the plot is always consistent with the table. For multi-model regression tables, attach a forest plot per model with .with_forest_plot() after slicing or call .with_forest_plot() on the merged table once.

Knobs:

Parameter	Effect
`log_x=True` (default)	Log-scale x-axis (natural for ORs / HRs / IRRs)
`null_line=1.0` (default)	Reference vertical line
`position='above'` / `'below'`	Plot placement relative to the table
`color='#0b3d91'`	Point + segment colour
`width_in=6.5`, `height_per_row_in=0.42`	Sizing

Kaplan–Meier curves¶

ps.tbl_survival(df, time='time', event='event', by='arm',
                times=[12, 24]).with_km_plot()

The KM curves are fit from the original data (the survival table preserves a reference) so they match exactly the medians and survival probabilities shown in the body rows.

Knobs:

Parameter	Effect
`ci=True` (default)	Shaded confidence bands
`palette=['#0b3d91', '#c1272d', ...]`	Per-group colours
`xlabel='Time'`, `ylabel='Survival probability'`	Axis labels
`width_in=6.5`, `height_in=4.0`	Figure size
`position='above'` / `'below'`	Plot placement

Custom SVGs¶

For any custom matplotlib output, render to SVG yourself and attach:

import io, matplotlib.pyplot as plt

fig, ax = plt.subplots()
# ... draw whatever ...
buf = io.StringIO()
fig.savefig(buf, format='svg', bbox_inches='tight')
my_svg = buf.getvalue()

table.with_inline_svg(my_svg, position='above')