fix(pptx): ignore empty llm captions

[Lubrsy (Lubrsy706)]

Summary

• Skip empty/None LLM image captions when building PPTX image alt text
• Preserve embedded deck alt text and shape-name fallback behavior
• Add regression coverage for an LLM caption provider returning None

Fixes #1534

Tests

• PYTHONPATH=packages/markitdown/src /Users/bytedance/MySpace/open-source/markitdown-excel-currency/.venv/bin/python -m pytest packages/markitdown/tests/test_module_misc.py::test_pptx_converter_treats_none_llm_caption_as_empty
• PYTHONPATH=packages/markitdown/src /Users/bytedance/MySpace/open-source/markitdown-excel-currency/.venv/bin/python -m compileall -q packages/markitdown/src/markitdown/converters/_pptx_converter.py packages/markitdown/tests/test_module_misc.py
• git diff --check

[Lubrsy (Lubrsy706)]

@microsoft-github-policy-service agree

[Copilot]

🟡 Changes recommended

Whitespace-only captions must preserve the shape-name fallback, and the regression assertion should verify the intended output.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Pull request overview

Prevents PPTX conversion failures when LLM image captioning returns None.

Changes:

• Filters missing LLM captions when composing image alt text.
• Adds regression coverage for None captions.

File summaries

File	Description
packages/markitdown/tests/test_module_misc.py	Adds regression coverage, but the assertion does not verify preservation of embedded alt text and uses the deprecated result alias.
packages/markitdown/src/markitdown/converters/_pptx_converter.py	Handles None captions, but whitespace-only captions can bypass the shape-name fallback.

Review details

• Files reviewed: 2/2 changed files
• Comments generated: 2
• Review effort level: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.