Empowering Data Science with Fast API Machine Learning Skeleton

• * ty documentation site and uv migration guide*
• * uv build backend is now stable + other Astral news*
• * Refactoring long boolean expressions*
• * fastapi-ml-skeleton*
• Extras
• Joke

About the show

Sponsored by Sentry: pythonbytes.fm/sentry

Connect with the hosts

• Michael: @mkennedy@fosstodon.org / @mkennedy.codes (bsky)
• Brian: @brianokken@fosstodon.org / @brianokken.bsky.social
• Show: @pythonbytes@fosstodon.org / @pythonbytes.fm (bsky)

Join us on YouTube at pythonbytes.fm/live to be part of the audience. Usually Monday at 10am PT. Older video versions available there too.

Finally, if you want an artisanal, hand-crafted digest of every week of the show notes in email form? Add your name and email to our friends of the show list, we'll never share it.

Michael #1: ty documentation site and uv migration guide

• via Skyler Kasko
• Astral created a documentation site for ty (PR #744 in release 0.0.1-alpha.13).
• Astral added a page on migrating from pip to a uv project in the uv documentation. (PR #12382 in release 0.7.19).
• Talk Python episode on ty.

Brian #2: uv build backend is now stable + other Astral news

• The uv build backend is now stable

  • Tim Hopper via Python Developer Tooling Handbook

• From Charlie Marsh

  • “The uv build backend is now stable, and considered ready for production use. An alternative to setuptools, hatchling, etc. for pure Python projects, with a focus on good defaults, user-friendly error messages, and performance. When used with uv, it's 10-35x faster.”

  • “(In a future release, we'll make this the default.)”

    [build-system]
    requires = ["uv_build>=0.7.19,<0.8.0"]
    build-backend = "uv_build"
    

  • I believe it’s faster, but I agree with Brett Cannon in asking “What's being benchmarked? I'm not sure what a "backend sync" is referring to other than maybe installing the build back-end?”

• See also: uv: Making Python Local Workflows FAST and BORING in 2025 - Hynek

Brian #3: Refactoring long boolean expressions

• Trey Hunner

• This is applied boolean logic, and even folks who learned this in a CS program probably did so early on, and may have forgotten it.

• How can you improve the readability of long Boolean expressions in Python?

  • Put parens around the whole expression and separate clauses onto different lines

  • Where to put boolean operators between clauses? at the end of the line or the beginning?

    • PEP8 recommends the beginning

      if (expression1
      and expression2
      and expression3):
      ...
      

  • Naming sub-expressions with variables

    • Odd downside that wouldn’t occur to me. All expressions are evaluated, thus not taking advantage of expression short-circuiting.

  • Naming operations with functions

    • Less readable, but takes advantage of short-circuiting

  • Using De Morgan’s Law : replacing a compound expression with a similar (and hopefully easier to read) expression

    # neither: we want both to be false
    not (a or b) == (not a) and (not b)
    
    # never_both: at least one false
    not (a and b) == (not a) or (not b)
    

Michael #4: fastapi-ml-skeleton

• FastAPI Skeleton App to serve machine learning models production-ready.
• This repository contains a skeleton app which can be used to speed-up your next machine learning project.
• The code is fully tested and provides a preconfigured tox to quickly expand this sample code.
• A sample regression model for house price prediction is included in this project.
• Short write up on "What does set -a do?"

Extras

Brian:

• OCF

Michael:

• via Wei Lee

• Extra Airflow ruff rules:

  Starting from Ruff version 0.11.13, most changes from Airflow 2 to Airflow 3 can be automated using AIR3. (It’s still in preview so a “—-preview” flag is needed)

  e.g., if you have the following Airflow 2 code

  import datetime
  
  from airflow.models import DAG
  from airflow.operators.empty import EmptyOperator
  
  with DAG(
    dag_id="my_dag_name",
    start_date=datetime.datetime(2021, 1, 1),
    schedule_interval="@daily",
  ):
  EmptyOperator(task_id="task")
  

  it can be fixed with uvx ruff check --select AIR3 --fix --unsafe-fixes --preview

  import datetime
  
  from airflow.sdk import DAG
  from airflow.providers.standard.operators.empty import EmptyOperator
  
  with DAG(
    dag_id="my_dag_name",
    start_date=datetime.datetime(2021, 1, 1),
    schedule="@daily",
  ):
  EmptyOperator(task_id="task")
  

  which works with Airflow 3.

Joke:

• Front Toward Enemy