Skip to main content

Writing a CLI

Introduction​

A CLI is helpful to run and debug code from the terminal. Flask-ML is able to automatically create a CLI for your machine learning code. Generating a CLI is a simple process that involves creating a single file.

Prerequisites​

  • At least one endpoint in your server file must be decorated with @server.route, and the route must have a task_schema_func parameter. See Adding a Task Schema for more information.
@server.route("/transform_case", task_schema_func=create_transform_case_task_schema)
def transform_case(inputs: TransformCaseInputs, parameters: TransformCaseParameters) -> ResponseBody:
    ...

Server File​

Ensure that your server file has a if __name__ == "__main__": block.

# server.py
ml_server = MLServer(__name__)

if __name__ == "__main__":
    ml_server.run()

CLI File​

Create a new file called cli.py in your project directory, and import your MLServer instance into it.

# cli.py

import argparse

from flask_ml.flask_ml_cli import MLCli
from server import ml_server


def main():
    parser = argparse.ArgumentParser(description="<Enter a description of your tool here>")
    cli = MLCli(ml_server, parser)
    cli.run_cli()


if __name__ == "__main__":
    main()

Running the CLI​

That's it! Now, you can run your CLI with:

python -m cli
$ python simple_cli.py --help
usage: simple_cli.py [-h] {transform_case} ...

Transform Case of multiple text inputs

positional arguments:
  {transform_case}  Subcommands
    transform_case  Transform Case

options:
  -h, --help        show this help message and exit

$ python simple_cli.py transform_case --help
usage: simple_cli.py transform_case [-h] --text_inputs TEXT_INPUTS [TEXT_INPUTS ...] [--to_case TO_CASE]

options:
  -h, --help            show this help message and exit
  --text_inputs TEXT_INPUTS [TEXT_INPUTS ...]
                        Text to Transform
  --to_case {upper,lower}
                        'upper' will convert all text to upper case. 'lower' will convert all text to lower case.