test: add poml format tests for vision prompts

[ultmaster]

Summary

• add tests covering dict, pydantic, openai_chat, and langchain outputs when poml renders image content

Testing

• npm run build-webview
• npm run build-cli
• npm run lint
• npm test
• python -m pytest python/tests

---

https://chatgpt.com/codex/tasks/task_e_689160949998832e8036fdadbc483dc0

[Copilot]

Pull Request Overview

This PR adds comprehensive test coverage for the POML (Prompt Markup Language) format functionality when rendering image content across different output formats. The tests verify that image content is correctly handled and formatted when converting POML markup to various API-specific formats.

• Adds tests for four output formats: dict, pydantic, openai_chat, and langchain
• Verifies correct image handling including base64 encoding, MIME type detection, and metadata preservation
• Ensures proper data structure formatting for each target API format

[Copilot]

The markup string is duplicated across all four test functions. Consider extracting this to a shared helper function or fixture to reduce code duplication and make the tests easier to maintain.

Suggested change

	def test_poml_format_dict(tmp_path: Path):
	img_path = _create_image(tmp_path)
	markup = f'<p>Image <img src="{img_path}" alt="tiny" syntax="multimedia"/></p>'
	def _make_markup(img_path: Path) -> str:
	return f'<p>Image <img src="{img_path}" alt="tiny" syntax="multimedia"/></p>'
	def test_poml_format_dict(tmp_path: Path):
	img_path = _create_image(tmp_path)
	markup = _make_markup(img_path)