The wizard of all wizards. WizPy aims to help developers turn classes and functions into a custom wizard, that walks the user through parameter gathering and execution.

This is a WIP - use at your own risk.

The wizpy is designed to be used with class definitions of an API. The API should be stateless, meaning it shouldn't be instantiated and thus regular methods on the class won't show in the wizard. The API class essentially will act as a thin wrapper around a group of functions, this allows the wizard to group functions into menus.

In the future, the wizard may be accepting of instantiated classes, however this was a design choice t keep the codebase consistent and opinionated, rather than implementation difficulties.

In must be said that both type annotations and succinct docstrings greatly improve the user experience when using the wizard, and as such are highly recommended through the codebase. To check typing, please use the mypy package or similar.

To register an API class to the wizard use the register_wizard decorator:

import boto3

from wizard import register_wizard


@register_wizard
class SageMakerAPI:

    @classmethod
    def list_endpoints(cls):
        cli = boto3.client('sagemaker')
        return cli.list_endpoints()

The register_wizard decorator can take a parameter that defines which functions in the class definition will be visible to the wizard by default, this is called description. If no parameters are passed as above then ALL functions will be visible

Please NOTE: that the API defined here only contains a classmethod. Only these methods can be used with wizard for now - this is a design decision to avoid having to instantiate and provide config to API instances.

If regular methods do indeed become required, such that the class needs to be instantiated, and gains enough support, then this can change and be raised as a feature.

This all works great, but what if the user of the wizard has no knowledge about the param they're suppose to pass?

Well a description can be added with the following decorator: add_parameter_description This decorator takes two args and one optional keyword

import boto3

from wizard import register_wizard


@register_wizard
class SageMakerAPI:

    @classmethod
    def list_endpoints(cls):
        cli = boto3.client('sagemaker')
        return cli.list_endpoints()

    @classmethod
    @add_parameter_description(
        parameter="endpoint_name",
        description="this is the endpoint_name: it requires a string"
    )
    def update_endpoint(cls, endpoint_name, endpoint_config_name):
        # do something

This will now add a description to the wizard as it instructs the user to input a value for this arg

As mentioned above, there is the option to pass in an optional keyword argument to this decorator: options This is to define what values the wizard can choose from, it can either be a list of values or a callback function.

import boto3

from wizard import register_wizard


def list_endpoint_config_names_view():
    cli = boto3.client('sagemaker')
    return cli.list_endpoint_configs()  # [config_1, config_2,... config_n]


@register_wizard
class SageMakerAPI:

    @classmethod
    def list_endpoints(cls):
        cli = boto3.client('sagemaker')
        return cli.list_endpoints()

    @classmethod
    @add_parameter_description(
        parameter="endpoint_config_name",
        description="this is the endpoint_name: it requires a string",
        options=list_endpoint_config_names_view

    )
    def update_endpoint(cls, endpoint_name, endpoint_config_name):
        # do something

NOTE: The callback function must return either a List, Dict or str.