Previously, we used the html2text crate, but now we have switched to html_to_markdown crate.

Here is the reason:

• html2text::from_read(html, wrap_width) requires a wrap_width, but for RAG, a larger wrap width is better. However, using usize::MAX can cause issues, as it may result in the generated markdown file being excessively large due to html2text automatically inserting spaces for wrap_width.
• html_to_markdown::convert_html_to_markdown allows us to control how each type of element is converted. For example, ignore <img>, <video> tags.