Dockerize a function and run it in a container directly from Python.
WARNING: pre-alpha state project, use only under your responsibility!
A way to run Python functions in a container without having to worry about docker.
- To keep some code with different dependencies than your main code.
- Because, why not?
You must have docker installed in your machine and docker python package (see requirements.txt file).
- Define a function that
- takes a file path as lone parameter,
- must have all imports locally, and
- will write the output in the file whose file path is the parameter.
- Use the functainerize decorator to run it on a container.
- 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.
- Check that the decorated function has the required parameter.
- Remove requirement to import everything locally.
- Add more tests.
- Add async version.
Diego J. Romero López