PyCon 2026 トークネタメモ

2026 トークネタメモイベント CfP

2026/03 PythonAsia 2026 のトークネタメモ

2026/05/15 PyCon US 2026 CfP応募したけど落ちた

2026/08/22 PyCon JP 2026 のトーク応募 link

提案1

Title

Creating Presentation Slides with the Retro Game Engine Pyxel

Description

Pyxel is a retro game engine developed by kitao-san, which is attractive because of its ability to create NES-like retro games in Python, its simple and intuitive API, and its expressive power limited to 16-colors and 4-sounds. On top of this strong Pyxel limitation, I created a presence-slide viewer that requires expressive power. I also incorporated an interesting mechanism using WebSocket communication. In this talk, I will show how I implemented the technology from loading Markdown to displaying slides, with a demonstration.

Combining existing libraries to achieve a goal within constraints is an everyday endeavor in modern programming.

In this talk, I will give you tips on how to deal with such efforts and details on how I specifically combined each library to create the presentation slides.

Pyxel: https://github.com/kitao/pyxel

Notes

This is a technical introduction to Pyxel Slide, implemented for PyCon mini Shizuoka 2024 Continue (Feb, 2025).

https://freia.jp/taka/slides/pyconshizu2024/

I wrote a blog post as a Pixel advent calendar 2025 which is also a draft presentation for this proposal.

https://freia.jp/taka/blog/2025/12/pyxel-slide/index.html

Outline

Introduction to Pyxel, a retro game engine (3min)

Demonstration of Pyxel presentation in action (2min)

Tokenize in markdown-it and output in visitor pattern (2min)

Rendering in Pyxel, character display in BDF font (2min)

Code highlighting in Pygments (2min)

Browser behavior with Pyodide + micropip (4min)

Multi-person operation via WebSocket communication (5min)

Category

Application of Python

Submit

PyCon US 2026

https://pretalx.com/pycon-us-2026/talk/review/MW9M8M8WMYR3GMSHDUM3DRPVYHDWPTZG

PythonAsia 2026

https://pretalx.com/python-asia-2026/talk/review/9L8NXVDH38FCFASMHDM3WMBFBJD8T39G

PyCon JP 2025

レトロゲームエンジンPyxelで作るプレゼンスライド :: PyCon JP 2025 :: pretalx

PyCon JP 2026

レトロゲームエンジンPyxelで作るプレゼンスライド :: PyCon JP 2026 :: pretalx

提案2

Title

10 ways to make pytest 10x faster!

Description

Do you write UnitTest? and how long do they take to run?

As projects grow, test suites get larger and slower.

Test runs that take a minute or two can discourage developers from running them often. Even when tests seem “fast enough,” making them 10× faster can dramatically change how frequently developers run them and act on the feedback.

In the PyPI case study, improved testing reduced the time from 163 seconds to 30 seconds, 80% reduction.

In my case study, it went from 351 seconds to 33 seconds, 90% reduction.

There are many advantages to having a test blast.

The obvious one is that you don't "have to wait" for the execution at hand, so you can easily rerun it.

It also eliminates the CI execution queue and reduces service fees.

This talk will cover 10 ways in which UnitTest execution time was reduced for real-world web service development code.

This talk will also introduce the concept of "what should be tested" and how to make decisions to replace bottlenecks. If we think, "In testing, we should reproduce the production behavior as much as possible," the same RDBMS, the same amount of data, the same settings, the same communication, etc., as in production will become bottlenecks, and time will accumulate. In order to replace them with another method, a decision must be made to replace them with a different method than the production method.

It is also important not to compromise the quality of testing for faster testing. 10 methods described in this article will speed up testing without compromising the level of quality assurance that was originally done. And when testing is accelerated, you will be able to test areas that were not covered in the past.

Notes

This talk is the English version of the content presented in Japanese at PyCon JP 2025.

It also builds upon the PyPI blog post “https://blog.trailofbits.com/2025/05/01/making-pypis-test-suite-81-faster/” and includes additional countermeasures.

Outline

1. Reduce setup time for database (3min)

2. Use pytest-xdist to run in parallel (3min)

3. Use sysmon for coverage (3min)

4. Squeeze pytest search paths (3min)

5. Stop unnecessary imports at test startup (3min)

6. Use pytest-profiling to find bottlenecks (3min)

7. Stop DB transaction testing (3min)

8. Narrow down the amount of data for testing (3min)

9. Use sqlite3 memory DB for DB (3min)

10. Find slowest test (3min)