Quick start guide

In this guide we will get started with PyFlyde by running a simple machine learning example that you can then modify and create your own projects in a similar way.

Copying an example project and installing dependencies

Download, copy, or clode the pyflyde repository.

Create a new directory for your project and copy the examples folder into it, e.g.

mkdir testflyde
cd testflyde
cp -R ../pyflyde/examples .

Edit the pyproject.toml file that comes with the examples adding pyflyde to the dependencies and removing the dev version of flyde package from it, so that it looks something like:

[project]
name = "pyflyde-examples"
version = "0.0.1"
requires-python = ">= 3.9"

dependencies = [
    "pyflyde >= 0.0.7",          # Add this line
    "matplotlib",
    "pandas",
    "scikit-learn",
]

[tool.setuptools.packages.find]
# where = ["../flyde", "mylib"]  # Replace this line
where = ["mylib"]                # With this line

Install the dependencies:

pip install examples/

Running the Hello World example

First, generate the component metadata for the examples:

pyflyde gen examples/

This will recursively scan all Python files in the examples/ directory and generate a flyde-nodes.json file with metadata for all PyFlyde components found.

Then run the example flow:

pyflyde examples/HelloPy.flyde

It should print "Hello Flyde!" in the console.

Running a Machine Learning example - wine clustering

examples/Clustering.flyde is a more complex example which uses Pandas and Scikit-Learn to run K-means clustering on a wine clustering dataset from Kaggle. It's a PyFlyde version of https://github.com/Shivangi0503/Wine_Clustering_KMeans.

The component metadata should already be generated from the previous step, but if you add new components, remember to run:

pyflyde gen examples/

Open the examples/Clustering.flyde in Flyde VSCode visual editor to see how it looks like.

To run this example, use the pyflyde command line tool:

pyflyde examples/Clustering.flyde

It should print dataframe clips for different steps of clusterization process and show a visualization at the end that looks like this: