feat(prompts): chat message placeholders

[nimarb]

Important

Adds support for chat message placeholders in langfuse-python SDK, enabling dynamic templating in prompts with comprehensive test coverage.

• Behavior:
  • Introduces ChatMessageWithPlaceholders type in chat_message_with_placeholders.py to handle both regular and placeholder messages.
  • Updates create_prompt_request.py and chat_prompt.py to use ChatMessageWithPlaceholders for prompt creation.
  • Adds PlaceholderMessage model in placeholder_message.py for template variables.
• Functionality:
  • Modifies ChatPromptClient in model.py to compile prompts with placeholders, handling unresolved placeholders by logging warnings.
  • Enhances get_langchain_prompt() to convert unresolved placeholders to MessagesPlaceholder objects.
• Testing:
  • Adds tests in test_prompt.py and test_prompt_compilation.py to cover placeholder scenarios, including edge cases and integration with Langchain.

^{This description was created by for 1c093d2. You can customize this summary. It will automatically update as commits are pushed.}

---

Greptile Summary

Disclaimer: Experimental PR review

Implements chat message placeholder functionality in the langfuse-python SDK, enabling dynamic templating in chat prompts. This feature is critical for scenarios like injecting conversation history or examples into prompts.

• Added ChatMessageWithPlaceholders union type in langfuse/api/resources/prompts/types/chat_message_with_placeholders.py supporting both regular chat messages and placeholder messages
• Modified prompt creation to use ChatMessageWithPlaceholders instead of ChatMessage in langfuse/api/resources/prompts/types/create_prompt_request.py
• Added new PlaceholderMessage Pydantic model in langfuse/api/resources/prompts/types/placeholder_message.py for template variables
• Added comprehensive test coverage in tests/test_prompt.py for placeholder scenarios and edge cases

[greptile-apps]

_{15 files reviewed, 1 comment
Edit PR Review Bot Settings | Greptile}

[ellipsis-dev]

Using hasattr(p, ...) in the prompt setter may fail for plain dict inputs which lack attribute access. Consider checking if p is a dict (e.g. using 'if isinstance(p, dict) and "type" in p') to support both dicts and objects.