Python News: What's New From May 2021

Flask Gets Native asyncio Support

According to the Python Developer Survey 2020, Flask is the most popular Python web framework. Flask 2.0’s native asyncio support is sure to keep the framework’s fans happy.

You can make everything from routes to error handlers to pre- and post-request functions into coroutines, which means you can define views using async def and await:

@app.route("/get-data")
async def get_data():
    data = await async_db_query(...)
    return jsonify(data)

In this example code snippet, taken from the Flask docs, an asynchronous view called get_data() is defined. It makes an asynchronous database query and then returns the data in JSON format.

Flask’s asyncio support is not without its caveats. Flask is still a Web Server Gateway Interface (WSGI) application and is subject to the same limitations as any other WSGI framework. Flask’s documentation describes these limitations:

Async functions require an event loop to run. Flask, as a WSGI application, uses one worker to handle one request/response cycle. When a request comes in to an async view, Flask will start an event loop in a thread, run the view function there, then return the result.

Each request still ties up one worker, even for async views. The upside is that you can run async code within a view, for example to make multiple concurrent database queries, HTTP requests to an external API, etc. However, the number of requests your application can handle at one time will remain the same. (Source)

If you’re new to asynchronous programming, check out Async IO in Python: A Complete Walkthrough. You can also get a breakdown of Flask’s new asyncio support from the article Async in Flask 2.0, which was featured in the PyCoder’s Weekly newsletter.

In addition to native asyncio support, Flask 2.0 also adds some new route decorators for common HTTP methods. For example, in Flask 1.x, you declare a route supporting the POST method using the @app.route() view decorator:

@app.route("/submit-form", methods=["POST"])
def submit_form():
    return handle_form_data(request.form)

In Flask 2.0, you can shorten this code using the @app.post() view decorator:

@app.post("/submit-form")
def submit_form():
    return handle_form_data(request.form)

That’s a small change but a significant improvement in readability!

You can find all the changes in Flask 2.0 in the official changelog.

Jinja Gets Improved Async Environments

asyncio support in Jinja 2.x required a patching system as well as some caveats developers needed to keep in mind. One of the reasons for this is that Jinja 2.x supports Python 2.7 and Python 3.5.

Now that all of the Pallets projects only support Python 3.6+, the patching system was removed to provide a more natural asyncio experience for projects using Jinja 3.0.

You can find all the changes in Jinja 3.0 in the official changelog.

Click Gets an Overhauled Shell Tab Completion System

Building great command-line interfaces (CLI) for an application can be a chore. The Click project helps ease this burden with its friendly API.

One of the features shell users expect from a CLI is tab completion, which suggests command names, option names, and choice values when a user types a few characters and presses Tab.

Click has always supported shell tab completion, but the implementation was messy, as Pallets maintainer David Lord noted on a March 2020 GitHub issue:

I’ve been trying to review [a pull request] which adds type-based completions, and it made me realize just how messy completion is, both within Click and with how the shells implement and document their systems.…

[W]e keep having to re-implement things that the shell should be doing, such as escaping special characters, adding spaces (except for directories), sorting, etc. And if the user wants to provide their own completions, they have to remember to do all that too.

There’s no reason we just have to return completions. We already support returning descriptions, presumably we could extend that more. Wouldn’t it be cool if Click could indicate to the completion script that it should use other functions that Bash or ZSH provide? (Source)

By October 2020, Click’s shell tab completion system had been completely overhauled with built-in support for Bash, Zsh, and fish. The system is extensible. You can add support for other shells, and the completion suggestions are customizable on multiple levels.

The new completion system, now available in Click 8.0, is a big win for projects that wish to offer users a friendly CLI experience in their favorite shell.

You can find a complete list of changes in Click 8.0 over on the official changelog.

A Virtual Conference That Feels Like the Real Thing

PyCon US 2020 was also virtual, but the last-minute transition to an online conference left the organizers little time to prepare an authentic PyCon experience. This year, the PSF had ample time to plan, and it delivered an incredibly engaging event true to the spirit of past PyCon US conferences.

Although talks were prerecorded, the videos were played on a schedule rather than being made available on demand. Each talk had a chat room associated with it, and presenters were available to interact with participants and answer questions.

The conference also featured a virtual expo hall that connected Pythonistas to various organizations in the Python world, including Microsoft, Google, Bloomberg, and more.

The most engaging part of PyCon 2021, though, was the well-executed open space and lounge area. Open spaces are small, meetup-like events that allow conference attendees to meet and interact with one another around a common interest. There were open spaces for Python authors, ham radio operators, community organizers, and so much more.

The conference’s lounge area included virtual tables that let a limited number of people join a videoconference. Anyone could grab an available chair and participate in the conversation, even if it had already started. The lounge gave PyCon a genuinely unique vibe, with all the spontaneity you’d expect at an in-person conference, effectively implementing a virtual hallway track, which is one of the hallmarks of PyCon US.

Moving PyCon US online made the conference accessible to more Pythonistas around the globe. No longer burdened by travel and accommodation expenses, Python enthusiasts could join PyCon from the comfort of their own homes for just the price of admission.

If you missed PyCon US 2021, you’ll soon be able to watch the talks on YouTube. The videos were still in postproduction at the time of writing but should be available in the coming weeks.

The Future of Python Is Focused on Performance

One of the goals of PyCon US is to bring together Python core developers and Python users to discuss the current state of and vision for the future of the language. Each year, the Python Language Summit brings together maintainers of Python implementations, such as CPython, PyPy, and Jython, to share information and solve problems.

This year’s Language Summit featured several exciting presentations. Dino Viehland talked about improvements Instagram made to CPython in its internal performance-oriented project called Cinder, including several enhancements to asynchronous I/O.

Python creator Guido van Rossum presented plans to make CPython faster. Van Rossum’s goal is to double CPython’s speed in time for Python 3.11. The increased performance would primarily benefit users who run CPU-intensive pure Python code or use tools and websites built in Python.

Another exciting feature of this year’s Python Language Summit is that the PSF gave Real Python’s own Joanna Jablonski the opportunity to cover the summit’s presentations and discussions in a series of blog posts. You can find all of her articles about the Language Summit on the PSF blog.