Fix br tail text handling in HTML tables

[dsolankii]

Fixes #3899.

Summary

This PR fixes text loss when normalizing HTML tables that contain <br/> tags inside table cells.

HtmlTable.from_html_text() previously removed all element tail text during normalization. For <br/> elements, the text after the tag is stored as tail text, so content after line breaks was dropped from the normalized HTML output.

Problem

Given an HTML table cell like:

This is 1st line.<br/>2nd line.<br/>3rd line.

the normalized table output preserved the <br/> tags but dropped the text after them, resulting in loss of 2nd line. and 3rd line..

Solution

This change preserves normalized tail text for <br/> elements while continuing to remove tail text for other elements.

This keeps the existing cleanup behavior for table normalization but avoids dropping valid content that follows line break tags.

Changes

• Preserve tail text for <br/> elements in HtmlTable.from_html_text().
• Continue removing tail text for non-br elements.
• Add regression coverage for preserving text after <br/> tags in HTML table cells.

How to test

Run the targeted regression test:

python -m pytest test_unstructured/common/test_html_table.py::DescribeHtmlTable::test_from_html_text_preserves_br_tail_text -q

Expected result:

1 passed

Validation

I reproduced the issue locally before applying the fix.

Before the fix, this input lost the text after <br/> tags:

This is 1st line.<br/>2nd line.<br/>3rd line.

After the fix, the normalized table preserves all lines:

This is 1st line.<br/>2nd line.<br/>3rd line.

The targeted regression test passes locally.

[dsolankii]

@cragwolfe Would you please review?