Plain text, and HTML that keeps the spans

Two more exports, and the choice between them is really a choice about what you are going to do with the result.

Plain text is less plain than it sounds

print(document.export_to_text())

The heading markers are gone, which is what plain means. The table is not gone: it is still a Markdown table, pipes and all, because there is no way to write a grid in plain text and Docling does not invent one.

So export_to_text is the right export when you want the words without the markup getting counted as words, and the wrong one if you were expecting the table to disappear.

HTML keeps everything Markdown dropped

from docling.document_converter import DocumentConverter

couriers = DocumentConverter().convert("couriers.html").document
print(couriers.tables[0].export_to_html(couriers))

There is the rowspan that lesson 10's Markdown had to flatten. HTML has the syntax for it, so the serializer uses it. When a merged header cell matters to whatever reads your output, this is the export to use.

A whole page, not a fragment

page = document.export_to_html()
print(len(page.splitlines()), "lines")
print(page.splitlines()[0])

export_to_html on the document writes a complete page with a stylesheet in it, ready to open in a browser. export_to_html on a single table, as above, writes just that table. Two methods with the same name at two levels, and the difference catches people out.

Choosing