Flaschenpost Challenge

Quick Start

Set up the environment

1. Install Poetry
2. Set up the environment:

make setup
make activate

If your IDE is unable to find this virtualenv in its 'select interpreter' command. Type the following in your shell:

poetry config virtualenvs.in-project true

Install new packages

To install new PyPI packages, run:

poetry add <package-name>

Get data

The data is stored in a gdrive folder.

To retrieve the data, simply type

dvc pull

to pull the data from the remote storage. That’s it!

Run Python scripts

To run the Python scripts to process data, train model, and run a notebook, type the following:

make pipeline

Create and run API

To create an API, type the following:

make api

We can now interact with the API by going to http://127.0.0.1:8000/docs and clicking the “Try it out” button:

You have to provide the image path. The image should live in this project. For example insert the following value:

{
  "img_path": "data/raw/fruits-360/Test/Apple Braeburn/3_100.jpg"
}

... to the Request body, which should give you "apple" as a response. This means the model correctly predicts that the image 3_100.jpg is an apple.

View all flow runs

A flow is the basis of all Prefect workflows.

To view your flow runs from a UI, sign in to your Prefect Cloud account or spin up a Prefect Orion server on your local machine:

prefect orion start

Open the URL http://127.0.0.1:4200/, and you should see the Prefect UI.

Run flows from the UI

After creating a deployment, you can run a flow from the UI with default parameters.

Auto-generate API documentation

To auto-generate API document for your project, run:

make docs_save

Run tests when creating a PR

When creating a PR, the tests in your tests folder will automatically run.