API Reference

promptpack

Functions

parse_promptpack(path: str | Path) -> PromptPack

Parse a PromptPack from a file path.

Arguments:

• path - Path to the PromptPack JSON file

Returns: PromptPack instance

Raises: PromptPackError on parse failure

parse_promptpack_from_dict(data: dict) -> PromptPack

Parse a PromptPack from a dictionary.

Arguments:

• data - Dictionary containing PromptPack data

Returns: PromptPack instance

Classes

PromptPack

Main class representing a parsed PromptPack.

Attributes:

• name: str - Pack name
• version: str - Pack version
• prompts: dict[str, Prompt] - Dictionary of prompts
• tools: list[Tool] - List of tool definitions
• fragments: dict[str, str] - Reusable fragments

Prompt

Represents a prompt definition.

Attributes:

• system_template: str | None - System message template
• user_template: str | None - User message template
• variables: dict[str, Variable] - Variable definitions

Methods:

• render(variables: dict) -> str - Render template with variables

Variable

Represents a template variable.

Attributes:

• type: str - Variable type (string, number, boolean, etc.)
• description: str | None - Variable description
• default: Any - Default value
• required: bool - Whether variable is required

Tool

Represents a tool definition.

Attributes:

• name: str - Tool name
• description: str - Tool description
• parameters: dict - JSON Schema for parameters

---

promptpack-langchain

Functions

convert_tools(pack: PromptPack) -> list[Tool]

Convert PromptPack tools to LangChain tool definitions.

Arguments:

• pack - PromptPack instance

Returns: List of LangChain tool definitions

Classes

PromptPackTemplate

LangChain prompt template wrapper for PromptPacks.

Class Methods:

• from_promptpack(pack: PromptPack, prompt_name: str) -> PromptPackTemplate

Attributes:

• chat_template: ChatPromptTemplate - Underlying LangChain template
• input_variables: list[str] - List of input variable names

Methods:

• format_messages(**kwargs) -> list[BaseMessage] - Format messages with variables

Validators

validate_variables(variables: dict[str, Variable], inputs: dict) -> list[str]

Validate input variables against definitions.

Arguments:

• variables - Variable definitions from PromptPack
• inputs - Input values to validate

Returns: List of validation error messages (empty if valid)

Multimodal

process_multimodal_content(content: dict) -> Any

Process multimodal content for LangChain compatibility.

Arguments:

• content - Content dictionary with type information

Returns: Processed content suitable for LangChain