Self Checks

• I have searched for existing issues, including closed ones.
• I confirm that I am using English to submit this report.

Describe the bug

In deepdoc/parser/html_parser.py, the TITLE_TAGS mapping currently defines:

TITLE_TAGS = {"h1": "#", "h2": "##", "h3": "###", "h4": "#####", "h5": "#####", "h6": "######"}

Here, both h4 and h5 are mapped to #####.

This seems incorrect. Based on Markdown heading levels, h4 should likely map to ####, while h5 should remain #####.

Expected behavior

TITLE_TAGS = {"h1": "#", "h2": "##", "h3": "###", "h4": "####", "h5": "#####", "h6": "######"}

Why this matters

When HTML headings are converted into text chunks, h4 content is currently promoted to the same level as h5, which loses heading hierarchy information and may affect downstream readability / chunk structure.

File

deepdoc/parser/html_parser.py

Suggested fix

Change:

"h4": "#####"

to:

"h4": "####"