Document VLM Captioning for Infographics

[kheiss-uwzoo]

Summary

Added comprehensive documentation explaining how to use VLMs to caption infographics across multiple documentation files.

The changes provide:

• Clear explanations of VLM captioning for infographics
• Code examples showing practical implementation
• Proper VLM acronym definitions
• Cross-references between related documentation sections

Changes Made

1. Added new section: "VLM Captioning for Infographics Example"

• Explains why infographics benefit from VLM captioning
• Provides three approaches:
• Approach 1: Extract and caption infographics (generates text descriptions)
• Approach 2: Embed infographics as images (preserves visual characteristics)
• Combining Both: Use both captioning and embedding together
• Includes complete code examples for each approach
• Added guidance on when to use each method

2. nv-ingest-python-api.md (Enhanced existing section)

• Enhanced "Extract Captions from Images" section to explicitly mention infographics
• Added new subsection: "Captioning Infographics"
• Includes code example showing extract_infographics=True with .caption()
• Added note about requiring the vlm profile
• Cross-referenced to vlm-embed.md

3. quickstart-guide.md (Enhanced profile documentation + example)

• Updated VLM profile description to mention infographics
• Added explanation that profile enables .caption() method
• Added new subsection: "Example: Using the VLM Profile for Infographic Captioning"
• Includes docker compose command with both retrieval and vlm profiles
• Provides complete end-to-end Python example
• Added tip explaining benefits of VLM captioning for complex visuals

4. support-matrix.md (Clarified VLM feature)

• Defined VLM acronym on first use (line 27)
• Updated VLM feature description to include infographics
• Ensures consistency across documentation

Documentation Standards

✅ Cross-references added between related sections
✅ Code examples are complete and runnable
✅ Progressive detail: quick reference → detailed examples → comprehensive guide
✅ Multiple entry points for users to discover this information

##Files Modified

• docs/docs/extraction/vlm-embed.md (+98 lines)
• docs/docs/extraction/nv-ingest-python-api.md (~30 lines modified/added)
• docs/docs/extraction/quickstart-guide.md (~50 lines modified/added)
• docs/docs/extraction/support-matrix.md (~1 line modified)

[nkmcalli]

A few suggestions for you