Test & Code

Many people have been working from home now that are not used to working from home. Or at least are working from home more than they ever did before. That's definitely true for me. Even though I've been working from home since March, I wanted some tips from people who have been doing it longer.Julian Sequeira, of PyBites fame, has been working from home for about a year. Reuven Lerner, an amazing Python trainer, has been working from home for much longer.We originally had a big list of WFH topics. But we had so much fun with the tips and tricks part, that that's pretty much the whole episode.But there's lots of great tips and tricks, so I'm glad we focused on that.Special Guests: Julian Sequeira and Reuven Lerner.Links:PyBites — Julian's site for teaching PythonTeaching Python and data science around the world — Reuven LernerBonbon - WikipediaTest & Code Mailing List — Join for your chance to win a free course from Talk Python Training. One course given away every week for 6 weeks.

Researches and others using data science and software need to follow solid software engineering practices. This is a message that Joel Grus has been promoting for some time.Joel joins the show this week to talk about data science, software engineering, and even Fizz Buzz.Topics include:Software Engineering practices and data scienceDifficulties with Jupyter notebooksCode reviews on experiment codeUnit tests on experiment codeFinding bugs before doing experimentsTests for data pipelinesTests for deep learning models Showing researchers the value of tests by showing the bugs found that wouldn't have been found without them."Data Science from Scratch" book Showing testing during teaching Data Science"Ten Essays on Fizz Buzz" book Meditations on Python, mathematics, science, engineerign and designTesting Fizz BuzzDifferent algorithms and solutions to an age old interview question.If not Fizz Buzz, what makes a decent coding interview question.pytesthypothesisMath requirements for data scienceSpecial Guest: Joel Grus.Links:Ten Essays on Fizz Buzz (with discount) by Joel GrusI don't like notebooks. (presentation)

pytest 6 is out. Specifically, 6.0.1, as of July 31. And there's lots to be excited about. Anthony Sottile joins the show to discuss features, improvements, documentation updates and more.Full release notes / changelogSome of what we talk about:How to update (at least, how I do it) Run your test suites with 5.4.3 or whatever the last version you were usingUpdate to 6Run again. Same output? Probably good.If there are any warnings, maybe fix those.You can also run with pytest -W error to turn warnings into errors.Then find out all the cool stuff you can do nowNew Features pytest now supports pyproject.toml files for configuration. but remember, toml syntax is different than ini files. mostly quotes are needed pytest now includes inline type annotations and exposes them to user programs. Most of the user-facing API is covered, as well as internal code.New command-line flags --no-header and --no-summaryA warning is now shown when an unknown key is read from a config INI file. The --strict-config flag has been added to treat these warnings as errors.New required_plugins configuration option allows the user to specify a list of plugins, including version information, that are required for pytest to run. An error is raised if any required plugins are not found when running pytest.Improvements You can now pass output to things like less and head that close the pipe passed to them. thank you!!!Improved precision of test durations measurement. use --durations=10 -vv to capture and show durationsRich comparison for dataclasses and attrs-classes is now recursive.pytest --version now displays just the pytest version, while pytest --version --version displays more verbose information including plugins. --junitxml now includes the exception cause in the message XML attribute for failures during setup and teardown.Improved Documentation Add a note about --strict and --strict-markers and the preference for the latter one.Explain indirect parametrization and markers for fixtures.Bug FixesDeprecationsTrivial/Internal ChangesBreaking Changes you might need to care about before upgrading PytestDeprecationWarning are now errors by default. Check the deprecations and removals page if you are curious.-k and -m internals were rewritten to stop using eval(), this results in a few slight changes but overall makes them much more consistenttestdir.run().parseoutcomes() now always returns the parsed nouns in plural form. I'd say that's an improvementSpecial Guest: Anthony Sottile.Links:pytest Changelog / Release NotesDeprecations and Removals — pytest documentation

pip is the package installer for Python. Often, when you run pip, especially the first time in a new virtual environment, you will see something like:WARNING: You are using pip version 20.1.1; however, version 20.2 is available. You should consider upgrading via the 'python -m pip install --upgrade pip' command.And you should. Because 20.2 has a new dependency resolver.Get in the habit, until October, of replacing pip install with pip install --use-feature=2020-resolver. This flag is new in the 20.2 release.This new pip dependency resolver is the result of a lot of work. Five of the people involved with this work are joining the show today: Bernard Tyers, Nicole Harris, Paul Moore, Pradyun Gedam, and Tzu-ping Chung. We talk about: pip dependency resolver changesuser experience research and testingcrafting good error messagesefforts to improve the test suitetesting pip with pytestsome of the difficulties with testing pipworking with a team on a large projectworking with a large code basebringing new developers into a large projectSpecial Guests: Bernard Tyers, Nicole Harris, Paul Moore, Pradyun Gedam, and Tzu-ping Chung.Links:Changelog — pip 20.2 documentation — Including --use-feature=2020-resolverpypa/pip: The Python package installer — github repotesting pip - documentationpip - The Python Package Installer — pip 20.2 documentationChanges to the pip dependency resolver in 20.2 — Changes to the pip dependency resolver in 20.2

Lots of Python projects are starting to use GitHub Actions for Continous Integration & Deployment (CI/CD), as well as other workflows.Tania Allard, a Senior Cloud Developer Advocate at Microsoft, joins the show to answer some of my questions regarding setting up a Python project to use Actions.Some of the topics covered:How to get started with GitHub Actions for a Python project?What are workflow files?Does it matter what the file name is called?Can I have / Should I have more than one workflow?Special Guest: Tania Allard.Links:Using Python with GitHub Actions - GitHub Docsawesome-actions — A curated list of awesome actions to use on GitHub

A great resume is key to landing a great software job. There's no surprise there. But so many people make mistakes on their resume that can very easily be fixed.Randall Kanna is on the show today to help us understand how to improve our resumes, and in turn, help us have better careers.Special Guest: Randall Kanna.Links:The Standout Developer — link includes discount

Len Wanger works on industrial 3D printers. And I was pleased to find out that there's a bunch of Python in those printers as well. In this episode we talk about:3D printersWhat are the different types of 3D printers?Where are 3D printed industrial parts being used?Why use one type of additive manufacturing over another?Python in 3D printing hardware.What are Finite State Machines, FSMs?Benefits of FSMs for testing, logging, and breaking a complex behavior into small testable parts.Benefits of simulation in writing and testing software to control hardware.Special Guest: Len Wanger.Links:pystate — Python package for co-routine base state machinesImpossible Objects — Composite 3D PrintingFinite-state machine, FSM

FastAPI is a modern, fast (high-performance), web framework for building APIs with Python based on standard Python type hints. Typer is a library for building CLI applications, also based on Python type hints. Type hints and many other details are intended to make it easier to develop, test, and debug applications using FastAPI and Typer.The person behind FastAPI and Typer is Sebastián Ramírez.Sebastián is on the show today, and we discuss:FastAPIRest APIsSwagger UIFuture features of FastAPIStarletteTyperClickTesting with Typer and ClickTyper autocompletionTyper CLISpecial Guest: Sebastián Ramírez.Links:ExplosionFastAPITyperOpenAPI Specification JSON SchemaOAuth 2.0StarlettepydanticSwagger UI — REST API Documentation ToolTesting - TyperClickTesting Click ApplicationsCLI Option autocompletion - TyperTyper CLI - completion for small scripts

There's stuff going on in Python packaging and pyproject.toml.Brett and I talk about some upcoming work on Python packaging, such as:editable installsthe need for standardizationconfiguration of other tools in pyproject.tomlAnd then get off on tangents and talk about:why it's good to have packages like pip, toml, setuptools, wheel, etc not part of the standard libraryshould we remove some stuff from the standard librarythe standard library using unittest for testing the standard library why not hypothesisI didn't bring up "why not pytest?" but you know I was thinking it.why CPython and not C++Pythonand moreSpecial Guest: Brett Cannon.Links:episode 52: pyproject.toml : the future of Python packaging - Brett CannonPython Packaging AuthorityPEP 517 -- A build-system independent format for source treesPEP 518 -- Specifying Minimum Build System Requirements for Python ProjectsWhat the heck is pyproject.toml?Flit PoetryensconstomlsetuptoolsdistutilspipHTTPX

Code Coverage or Test Coverage is a way to measure what lines of code and branches in your code that are utilized during testing.  Coverage tools are an important part of software engineering. But there's also lots of different opinions about using it. Should you try for 100% coverage? What code can and should you exclude? What about targets?I've been asked many times what I think about code coverage or test coverage. This episode is a train of thought brain dump on what I think about code coverage.We'll talk about:how I use code coverage to help me write source codeline coverage and branch coveragebehavior coverageusing tests to ask and answer questions about the system under testhow to target coverage just to the code you care aboutexcluding codegood reasons and bad reasons to exclude codeAnd also the Pareto Principle or 80/20 rule, and the law of diminishing returns and how that applies (or doesn't) to test coverage.Links:Coverage.pypytest-cov