mirascope.core.vertex.tool

The VertexTool class for easy tool usage with Google's Vertex LLM calls.

Usage Documentation

Tools

VertexTool

Bases: BaseTool

A class for defining tools for Vertex LLM calls.

Example:

from mirascope.core import prompt_template
from mirascope.core.vertex import vertex_call


def format_book(title: str, author: str) -> str:
    return f"{title} by {author}"


@vertex_call("gemini-1.5-flash", tools=[format_book])
def recommend_book(genre: str) -> str:
    return f"Recommend a {genre} book"


response = recommend_book("fantasy")
if tool := response.tool:  # returns an `VertexTool` instance
    print(tool.call())

tool_schema classmethod

tool_schema() -> Tool

Constructs a JSON Schema tool schema from the BaseModel schema defined.

Example:

from mirascope.core.vertex import VertexTool


def format_book(title: str, author: str) -> str:
    return f"{title} by {author}"


tool_type = VertexTool.type_from_fn(format_book)
print(tool_type.tool_schema())  # prints the Vertex-specific tool schema

Source code in mirascope/core/vertex/tool.py

@classmethod
def tool_schema(cls) -> Tool:
    """Constructs a JSON Schema tool schema from the `BaseModel` schema defined.

    Example:
    ```python
    from mirascope.core.vertex import VertexTool


    def format_book(title: str, author: str) -> str:
        return f"{title} by {author}"


    tool_type = VertexTool.type_from_fn(format_book)
    print(tool_type.tool_schema())  # prints the Vertex-specific tool schema
    ```
    """
    model_schema = cls.model_json_schema()
    fn: dict[str, Any] = {"name": cls._name(), "description": cls._description()}
    if model_schema["properties"]:
        fn["parameters"] = model_schema
    if model_schema["required"]:
        fn["parameters"]["required"] = model_schema["required"]
    if "parameters" in fn:
        if "$defs" in fn["parameters"]:
            raise ValueError(
                "Unfortunately Google's Vertex API cannot handle nested structures "
                "with $defs."
            )

        def handle_enum_schema(prop_schema: dict[str, Any]) -> dict[str, Any]:
            if "enum" in prop_schema:
                prop_schema["format"] = "enum"
            return prop_schema

        fn["parameters"]["properties"] = {
            prop: {
                key: value
                for key, value in handle_enum_schema(prop_schema).items()
                if key != "default"
            }
            for prop, prop_schema in fn["parameters"]["properties"].items()
        }
    return Tool(function_declarations=[FunctionDeclaration(**fn)])

from_tool_call classmethod

from_tool_call(tool_call: FunctionCall) -> VertexTool

Constructs an VertexTool instance from a tool_call.

Parameters:

Name	Type	Description	Default
tool_call	FunctionCall	The Vertex tool call from which to construct this tool instance.	required

Source code in mirascope/core/vertex/tool.py

@classmethod
def from_tool_call(cls, tool_call: FunctionCall) -> VertexTool:
    """Constructs an `VertexTool` instance from a `tool_call`.

    Args:
        tool_call: The Vertex tool call from which to construct this tool instance.
    """
    if not tool_call.args:
        raise ValueError("Tool call doesn't have any arguments.")
    model_json: dict[str, Any] = dict(tool_call.args.items())
    model_json["tool_call"] = tool_call
    return cls.model_validate(model_json)