20190505_pycon_plugins

Plug-ins: Adding Flexibility to Your Apps

Talk given at PyCon, Cleveland, OH, USA on May 5, 2019

• [Abstract]
• [Slides]
• [Video]

The slides were presented using mdp. The content of the slides are available below.

The code is available in two different directories:

• code_talk contains the code as written during the presentation
• code_pyplugs contains the expanded code based on pyplugs shown at the end of the talk

The pyplugs library is available at PyPI.

Presentation

This is the content of the presentation slides, adapted to a flat text file:

Plug-ins: Adding Flexibility to Your Apps

Geir Arne Hjelle

---

Agenda

• Motivation

• Example: Plotter app

  • Different data formats
  • Choosing what to plot
  • More plotting styles

• Plug-ins

---

Motivation

Modularized code:

• is less complex
• is easier to maintain and test
• is easier to extend

With plug-ins, your app:

• can be controlled by configuration settings
• can be extended and customized for and by users
• can be better structured, and developed faster

---

Demo time

We'll build a simple plotter app:

• Command line application
• Read data from a CSV file
• Plot data in a simple line graph

---

Decorators

A decorator is a function that wraps another function:

decorated = decorator(original)

• Decorators are typically used to add some common functionality across many functions or methods

• In most cases, decorators are applied using @-syntax:

  @click.command()
  def main(file_path): 
      ...
  

realpython.com/primer-on-python-decorators

---

Demo time

Let's expand our plotter app:

• Support more file formats, like JSON
• Support more kinds of plots
• Control which data to plot

---

PyPlugs

A simple decorator based plug-in architecture for Python

$ pip install pyplugs

Three levels:

• Plug-in packages: Directories containing files with plug-ins
• Plug-ins: Modules containing registered functions or classes
• Plug-in funcs: Several registered functions in the same file

---

PyPlugs

Example use cases for plug-ins:

• Readers for different file formats
• Models for flexible calculations
• Filters for filtering your data
• Writers for storing your data in different formats
• Notifiers for sending you data to different devices
• Components for building your GUI

---

Thank You For Your Attention

• Me: @gahjelle
• Code: github.com/gahjelle/talks
• PyPlugs: pyplugs.readthedocs.io
• Real Python: realpython.com