Alternatives, Inspiration and Comparisons
What inspired Typer, how it compares to other alternatives and what it learned from them.
Intro¶
Typer wouldn't exist if not for the previous work of others.
There have been many tools created before that have helped inspire its creation.
Previous tools¶
argparse¶
argparse is the Python standard library's module to write CLIs.
It provides a better alternative than reading the CLI Parameters as a list of str and parsing everything by hand.
Inspired Typer to
Provide a better development experience than just reading CLI Parameters by hand.
Hug¶
Hug is a library to create APIs and CLIs, it uses parameters in functions to declare the required data.
It inspired a lot of the ideas in FastAPI and Typer.
Inspired Typer to
Use function parameters to declare CLI arguments and CLI options as it simplifies a lot the development experience.
Plac¶
Plac is another library to create CLIs using parameters in functions, similar to Hug.
Inspired Typer to
Provide a simple way to use a function as a command-line app, without having to create a complete app, with typer.run(some_function).
Pydantic¶
Pydantic is a library to handle data validation using standard modern Python type annotations.
It powers FastAPI underneath.
It is not used by Typer, but it inspired a lot of the design (through FastAPI).
Inspired Typer to
Use standard Python type annotations to declare types instead of library-specific types or classes and use them for data validation and documentation.
Click¶
Click is one of the most widely used libraries to create CLIs in Python.
It's a very powerful tool and there are many CLIs built with it. It is what powers Typer underneath.
It also uses functions with parameters for CLI arguments and CLI options, but the declaration of the specific CLI arguments, CLI options, types, etc, is done in decorators on top of the function. This requires some code repetition (e.g. a CLI Option name --verbose and a variable name verbose) and synchronization between two places related to the same information (the decorator and the parameter function).
It uses decorators on top of functions to modify the actual value of those functions, converting them to instances of a specific class. This is a clever trick, but code editors can't provide great support for autocompletion that way.
It was built with some great ideas and design using the features available in the language at the time (Python 2.x).
Typer uses it for
Everything. 🚀
Typer mainly adds a layer on top of Click, making the code simpler and easier to use, with autocompletion everywhere, etc, but providing all the powerful features of Click underneath.
As someone pointed out: "Nice to see it is built on Click but adds the type stuff. Me gusta!"
FastAPI¶
I created FastAPI to provide an easy way to build APIs with autocompletion for everything in the code (and some other features).
Typer is the "FastAPI of CLIs".
It uses the same design and usage of FastAPI as much as possible. So, if you have used FastAPI, you know how to use Typer.