First steps

To start this example you should first create a virtual env.

$python3 -m venv ./venv_example
$source ./venv_example/bin/activate                        (Linux)
$.\venv_example\Scripts\activate.bat                 (Windows cmd)

After the venv has been created you should install the requirements

$pip install -r requirements.txt
$pip install -e .

The script model_training can be used to train and serialise a scikit-learn model

$python ./model/model_training.py
$git add ./model.pkl
$git commit -m 'aggiunto modello'
$git push

Definition of a unit test

The file test/test_data_and_model.py contains a pytest example. Local tests can be run by using the command

$python -m pytest

At each execution pytest collects and runs every function containing the word 'test' in its name.

Performing a grid-search with mlflow

The script run_grid_search contains the logic to perform a grid search using mlflow to store the results

$python experiments/run_grid_search.py

You can access the mlflow user interface by running

$mlflow ui

You can use the query bar to search for specific experiments, for example

metrics.precision > 0.6 and params.depth='3'