PyCon 2019 Talks I've Watched

Mocking and Patching Pitfalls

• Slides
• Nine levels of hell, leading to tech debt
• References and builds on prior talks and writing
• Puts mocking in historical TDD/OOP context, disambiguates it from other kinds of test doubles
• Works through example of getting out of hell via stubs/fakes, seams, dependency injection, clean architecture, etc.

Wily Python: Writing simpler and more maintainable Python

• Slides
• Good overview of utility/math of cyclomatic complexity, Halsted metrics, and maintainability index
• wily graphs metrics trends based on git history
• Fewer lines != less complex, Test coverage != behavior coverage

The Zen of Python Teams

• Interprets six lines of "Zen of Python" as they apply to team culture

Assets in Django without losing your hair

• High-level overview of simple-to-complex scenarios
• Start simple with WhiteNoise, eventually use django-storages for media on S3 and django-webpack-loader for modern JavaScript

Break the Cycle: Three excellent Python tools to automate repetitive tasks

• Slides
• Fun problem statement and narrative about reducing toil with automation
• tox for testing, managing virtual environments, via .ini files
• nox for same, plus general project automation, via Python
• invoke for general task-based automation, via Python

Building a Culture of Observability

• Slides
• Observability == What's going on? Why did things go wrong?
• How to improve logging and metrics to tell the narrative of a web app

Intentional Deployment: Best Practices for Feature Flag Management

• How to manage the complexity and tech debt that arise from feature flags
• How to adapt company processes and culture around feature flags

The Perils of Inheritance: Why We Should Prefer Composition

• Slides
• Balanced view of a common debate, with examples and pros/cons of both approaches

PEP 572: The Walrus Operator

• Slides
• Good overview of why, use cases of Python 3.8 feature
  • [y for x in data if y := f(x)]
  • if match := pattern.search(data)
• Good overview of PEP process, controversy, steering council

Syntax Trees and Python - Automated Code Transformations

• Good introduction to theory of abstract/concrete syntax trees, with examples
• Used in lib2to3, futurize, black, etc.
• Also useful for refactoring (but not clear how)

Life Is Better Painted Black, or: How to Stop Worrying and Embrace Auto-Formatting

• Reasoning and implementation behind the popular "Opinionated, Deterministic, Consistent, Fast" formatter from its creator
• Use automation (instead of noisy linter) to focus on logic, not code style
• Builds on syntax trees talk

Type hinting (and mypy)

• Slides
• Blog post: the state of type hints in Python
• Why and how, with current gotchas and coming improvements
• Easier understandability, maintainability, editing; like unit tests, but for type correctness

Getting to Three Million Lines of Type-Annotated Python

• Narrative about Dropbox's motivation for and challenges faced during adoption
• Annotate legacy code to support new code and enable "aggressive refactoring"
• Used social pressure for adoption
• Performance challenges impeding adoption (and their solutions) were due to scale

The Refactoring Balance Beam: When to Make Changes and When to Leave it Alone

• Slides
• A little too abstract/simple/quick, but still a decent introduction to changing code structure, not behavior

Nina Zakharenko Keynote

• Slides
• CircuitPython demo: make blinky lights and noise with Python!
• Lots of links in slides PDF

Don't be a robot, build the bot

• Slides
• Nice walkthrough from problem to implementation of miss-islington, a bot for backporting CPython pull requests

Shipping your first Python package and automating future publishing

• Slides
• Generally follows official PyPA guide to publish your first package so you can pip install
• Example of automation via nox and CircleCI

The Black Magic of Python Wheels

• Elana Hashman
• Talk Resources
• Decent background on evolution of wheels
• Most focus on native extensions (e.g. cryptography library) and manylinux effort to improve adoption

Escape from auto-manual testing with Hypothesis!

• Slides
• Lot of theory, not an introduction; there are other docs and talks for that