simpletool 0.0.4

simpletool

Simpletool

A Python package for creating simple tool class with a base class and data type definitions.

Overview

This package provides a BaseTool class that serves as the foundation for building various tools. It includes functionalities for:

• Executing tools with input arguments.
• Handling environment variables.
• Generating JSON schemas for input models.

The package also defines data types for content, resources, and errors, ensuring consistency and clarity in tool development.

Core Components

BaseTool Class

The BaseTool class is an abstract base class that all tools should inherit from. It provides the following methods:

• run(arguments): Executes the tool with the given arguments. This method should be overridden by subclasses to implement the tool's logic.
• execute(arguments): An alternative name for the run method.
• get_env(arguments, prefix): Retrieves environment variables from arguments, os.environ, and resources, optionally filtering by a prefix.
• to_json(input_model, schema): Converts an input model to a JSON schema.

Data Types

The types.py file defines the following data types:

• Content: Base class for content types.
• TextContent: Represents text content.
• ImageContent: Represents image content.
• ResourceContents: Represents the contents of a resource.
• TextResourceContents: Represents the contents of a text resource.
• BlobResourceContents: Represents the contents of a blob resource.
• EmbeddedResource: Represents an embedded resource.
• ErrorData: Represents error information.

Installation

To install the package, use pip:

pip install simpletool

Usage

To create a new tool, inherit from the BaseTool class and implement the run or execute method. The input schema can be automatically generated from a Pydantic model using the to_json method.

Example 1: Input schema as a dictionary

from simpletool import BaseTool, TextContent
from typing import Dict, Any, List, Union

class MyTool(BaseTool):
    name = "my_tool"
    description = "A simple example tool"
    input_schema = {
        "type": "object",
        "properties": {
            "message": {
                "type": "string",
                "description": "The message to print"
            }
        },
        "required": ["message"]
    }

    async def run(self, arguments: Dict[str, Any]) -> Union[List[TextContent], ErrorData]:
        message = arguments["message"]
        return [TextContent(type="text", text=f"You said: {message}")]

Example 2: Input schema as a Pydantic model

from simpletool import BaseTool, TextContent, NoTitleDescriptionJsonSchema
from typing import Dict, Any, List, Union
from pydantic import BaseModel, Field

class InputModel(BaseModel):
    """Input model for time conversion."""
    date_time_str: str = Field(
        description="The time to convert. Can be 'NOW' or a specific date and time in a format like 'YYYY-MM-DD HH:MM:SS'."
    )
    from_timezone: str = Field(
        default="",
        description="Source timezone (default: local timezone)"
    )
    to_timezone: str = Field(
        default="UTC", 
        description="Target timezone (default: UTC)"
    )
args_schema = InputModel.model_json_schema(schema_generator=NoTitleDescriptionJsonSchema)

class MyTool2(BaseTool):
    name = "my_tool2"
    description = "A second simple example tool"
    input_schema = InputModel2.model_json_schema(schema_generator=NoTitleDescriptionJsonSchema)

class InputModel2(BaseModel):
    """Input model for another tool."""
    name: str = Field(description="The name of the user.")
    age: int = Field(description="The age of the user.")

    async def run(self, arguments: Dict[str, Any]) -> Union[List[TextContent], ErrorData]:
        message = arguments["message"]
        return [TextContent(type="text", text=f"You said: {message}")]

Accessing Tool Information

The BaseTool class provides several ways to access tool information:

• str(my_tool): Returns a one-line JSON string representation of the tool.
• my_tool.details: Returns a one-line JSON string representation of the tool.
• my_tool.to_dict: Returns a dictionary representation of the tool. Note that my_tool.__dict__ is not the same as would return a dictionary containing all attributes of the object, including internal ones.

from simpletool import BaseTool, NoTitleDescriptionJsonSchema
from simpletool.types import TextContent, ImageContent, EmbeddedResource, ErrorData
from pydantic import BaseModel, Field
from typing import List, Union


class InputSchema(BaseModel):
    """Input model for the MyTool"""
    message: str = Field(description="The message to print")


class MyTool(BaseTool):
    name = "my_tool"
    description = "A simple example tool"
    input_schema = InputSchema.model_json_schema(schema_generator=NoTitleDescriptionJsonSchema)

    async def run(self, arguments: dict) -> Union[List[TextContent | ImageContent | EmbeddedResource], ErrorData]:
        """Execute the tool with the given arguments"""
        # Validate input using Pydantic model
        input_model = InputSchema(**arguments)
        message = input_model.message
        return [TextContent(type="text", text=f"You said: {message}")]


my_tool = MyTool()

print("\nString Representation - str(my_tool):")
print(f"Type: {type(str(my_tool))}")
print(str(my_tool))
# output:
#String Representation - str(my_tool):
#Type: <class 'str'>
#{"name": "my_tool", "description": "A simple example tool", "input_schema": {"properties": {"message": {"type": "string"}}, "required": ["message"], "type": "object"}}

print("\nTool Details - my_tool.info:")
print(f"Type: {type(my_tool.info)}")
print(my_tool.info)
# output:
#Tool Details - my_tool.info:
#Type: <class 'str'>
#{
#    "name": "my_tool",
#    "description": "A simple example tool",
#    "input_schema": {
#        "properties": {
#            "message": {
#                "type": "string"
#            }
#        },
#        "required": [
#            "message"
#        ],
#        "type": "object"
#    }
#}

print("\nDictionary Representation - my_tool.to_dict:")
print(f"Type: {type(my_tool.to_dict)}")
print(my_tool.to_dict)
# output:
#Dictionary Representation - my_tool.to_dict:
#Type: <class 'dict'>
#{'name': 'my_tool', 'description': 'A simple example tool', 'input_schema': {'properties': {'message': {'type': 'string'}}, 'required': ['message'], 'type': 'object'}}

Dependencies

• pydantic>=2.0.0
• typing-extensions

License

This project is licensed under the MIT License.

Contributing

Contributions are welcome! Please submit a pull request with your changes.