Exports

Every SofraTable consumes the same internal schema, so all renderers produce consistent output.

Method	Format	Notes
.to_html(sticky_header=False, max_height=None)	HTML fragment	Scoped CSS, inherit colours
.to_markdown()	GitHub-flavored Markdown	Alignment markers preserved
.to_docx(path)	Word .docx	Publication-quality, python-docx fully abstracted
.to_latex(booktabs=True, float_position='ht', centering=True)	LaTeX string	Booktabs by default; falls back to \hline
.to_pptx(path, slide_title=None)	PowerPoint .pptx	Optional extra pysofra[pptx]
.to_xlsx(path, sheet_name='Table')	Excel .xlsx	Numeric cells written as numbers; captions + footnotes + spanning headers preserved

Notebook reprs

SofraTable exposes:

• _repr_html_ — used by Jupyter, Colab, VS Code, Quarto
• _repr_markdown_ — used by Markdown-first viewers
• _repr_latex_ — used by LaTeX-first frontends

Frontends try _repr_html_ first by default, so the rich notebook render is what you see.

Round-tripping

The internal SofraTable is the source of truth. Exporting the same object to HTML and DOCX produces consistent content (text, alignment, captions, footnotes, spanning headers). Theme decisions that are HTML-specific (CSS classes, color-mix) degrade naturally to DOCX (font_name, font_size, border hints).