diegojromerolopez / functainer

Dockerize a function and run it in a container

functainer

Dockerize a function and run it in a container directly from Python.

WARNING: pre-alpha state project, use only under your responsibility!

What's this?

A way to run Python functions in a container without having to worry about docker.

Why?

• To keep some code with different dependencies than your main code.
• Because, why not?

Requirements

You must have docker installed in your machine and docker python package (see requirements.txt file).

Instructions

1. Define a function that
   1. takes a file path as lone parameter,
   2. must have all imports locally, and
   3. will write the output in the file whose file path is the parameter.
2. Use the functainerize decorator to run it on a container.
3. That's all.

@functainerize(image='python:3.10', requirements=['pandas==1.5.0'])
def package_installation(output_file_path: str):
    import pandas as pd

    url = "https://raw.githubusercontent.com/cs109/2014_data/master/countries.csv"
    df = pd.read_csv(url)
    
    # Do some operations with the dataset

    df.to_pickle(path=output_file_path)

For more examples, see the test_decorator.py file.

TODO

• Check that the decorated function has the required parameter.
• Remove requirement to import everything locally.
• Add more tests.
• Add async version.

Author

Diego J. Romero López

License

MIT