Build a quick inference CLI for your machine learning model

Convert your machine learning python script to a quick CLI tool in a modern python style using fire and poetry

Build a quick inference CLI for your machine learning model
Photo by Daphne on Unsplash

Convert your machine learning python script to a quick CLI tool in a modern python style using fire and poetry

Here we are going to quickly look at the python-fire package from google to bootstrap your python script into a CLI tool.

Python Fire is a library for automatically generating command-line interfaces (CLIs) from absolutely any Python object.”- Google

You can find all the quick examples from this Github repo: https://github.com/google/python-fire

but here we will look into a few extended use cases of fire.


Quick Usage

Create a python file and start with a simple script:# myscript.pyfrom typing import Unionfrom somewhere import modelimport fire


def predict(height: Union[int, float], weight: Union[int, float]) -> Union[int, float]:
return model.predict([height, weight])


if __name__ == "__main__":
fire.Fire(add)

Now, if you can run this file as a CLI using -m flag and passing 2 arguments in python.python -m myscript 155 60

Output
MALE

that was pretty quick and easy.

Class example

You can also create a CLI from the python class. Here is an example:import fire
from somewhere import model

class Predict:
@staticmethod
def gender(weight, model):
return model.predict_gender(weight, model)

@staticmethod
def age(weight, b):
return model.predict_age(weight, model)

@staticmethod
def BMI(weight, b):
return a / b

@staticmethod
def shoe_size(weight, height):
return model.get_shoe_size(weight, height)


if __name__ == "__main__":
fire.Fire(Predict)

We now have all those different methods we want in our calculator. We can run this using the following command:

  1. For age prediction:python -m myscript age 155 60Output
    30

2. for shoe size prediction:

python -m myscript
Output 
44

And so on….

Poetry Integration

Poetry is a package manager for python similar to npm for NodeJS.

Poetry deserves a separate story, but if you want to learn more from the docs, here it is: https://python-poetry.org/.

We can integrate fire CLI with poetry for easy scripting. Here is a way to do that:

  1. Create a script and a runner file.
  2. In the runner file import the script that you want to run and put the name-main guard.# runner.py
    from myscript import Predictdef main():
    fire.Fire(Predict)


    if __name__ == "__main__":
    main()

3. Add an entry point in your `pyproject.toml` file below your dependencies.[tool.poetry.scripts]
predict = "runner:main"

4. Install your packages

poetry install

5. Now you can run your script simply using this command:predict age 155 60

Output 
30

As easy as that 😀.

If this post was helpful, please click the clap 👏 button below a few times to show your support for the author 👇

🚀Developers: Learn and grow by keeping up with what matters, JOIN FAUN.

Subscribe to CodeCraft by Anyesh | Programming, ML & AI Tutorials

Don’t miss out on the latest issues. Sign up now to get access to the library of members-only issues.
jamie@example.com
Subscribe