Is there a way to get section wise HTMLs from SEC document?

[ushnish]

I am new to using this library. I know that there is a way to get the full HTML for a filing and there’s a way to get each item’s section from a filing as plain text, but is there a way to get each section as full HTML? Like the entire section for “Item 1. Business” as one single html. Thank you in advance

[dgunning]

Technical Analysis

Thank you for this excellent question! You've identified an important gap in EdgarTools' current capabilities. Let me provide a comprehensive answer based on our analysis of the codebase.

Current State

You're absolutely correct in your understanding. EdgarTools currently provides:

1. Full HTML access via filing.html() - returns the complete filing HTML
2. Plain text sections via tenk['Item 1'] - extracts individual sections as text

filing = ...
tenk = filing.obj()
item1 = tenk["Item 1"]

4. Markdown conversion via filing.markdown() - converts to markdown format

However, there is no direct method to extract individual sections as HTML fragments. This is indeed a missing feature that would be valuable for many use cases.

Implementation Considerations

After analyzing the codebase, here's what we found:

Why This Doesn't Currently Exist

The current architecture in ChunkedDocument (which powers section extraction) was designed with a text-first approach:

• When parsing, the HTML is broken into blocks (text, tables, links)
• These blocks are stored in a simplified format
• The original HTML structure is discarded during parsing
• Section assembly methods exist only for text and markdown output

Technical Feasibility

The good news is that this feature is technically feasible. We've identified several potential implementation approaches:

1. Extend existing ChunkedDocument - Add HTML preservation during parsing
2. Create parallel HTML-aware parser - New implementation that maintains HTML structure
3. Use anchor-based extraction - Modern filings have section anchors we could leverage

Community Questions

To help shape this potential feature, I'd love to understand your use case better:

1. What would you do with section HTML?

   • Display it in a web interface?
   • Further parse it for specific elements?
   • Transform it into another format?

2. HTML cleanup preferences?

   • Should we include inline styles and CSS?
   • Keep or remove embedded XBRL tags?
   • Preserve or simplify table formatting?

3. API design preference?

   • New method like tenk.get_item_html('Item 1')?
   • Format parameter like tenk.get_item('Item 1', format='html')?
   • Property access like tenk.item1_html?

Next Steps

This is clearly a valuable feature that aligns with EdgarTools' mission of providing powerful yet simple APIs. I'll document this as a potential enhancement. Your feedback on the questions above would help shape the implementation to best serve the community's needs.

Would you be interested in testing a prototype implementation if we develop one? Your real-world use case would be invaluable for ensuring we get the API design right.

Thank you for bringing this up - it's exactly the kind of feedback that helps make EdgarTools better for everyone!