Welcome to issue 682 of Python Weekly. Happy New Year! I hope you had a great holiday and took some time off to recharge.

Learn how to make AI work for you

AI won’t take your job, but a person using AI might. That’s why 800,000+ professionals read The Rundown AI – the free newsletter that keeps you updated on the latest AI news and teaches you how to use it in just 5 minutes a day.

Sign up to start learning.

Articles, Tutorials and Talks

How I Built a Full-Stack AI App in 5 Hours (Agents, Streamlit, Python & More)  

The video showcases the process of building an AI coding tutor/assistant using Parlant, a framework for guided LLM agents. The project includes features like roadmap generation, resource tracking, and quiz creation, demonstrating the integration of AI tools with a Python backend and Streamlit frontend.

Can LLMs write better code if you keep asking them to “write better code”? 

The article explores iteratively asking an LLM to "write better code" for a Python task, resulting in significant performance improvements through algorithmic optimizations and parallelization. While initial iterations yielded substantial speedups, later iterations led to overengineered "enterprise-level" features, highlighting the importance of specific prompting for desired outcomes in LLM-assisted coding.

Debugging FastAPI in VSCode Is Easier Than You Think

Debugging a FastAPI app in VSCode is tricky because the code is run via a tool like uvicorn , and you typically interact with the API via curl statements or a browser. This video shows you how to setup debugging for your FastAPI app.

Symbolic Execution by Overloading __bool__

The article discusses a technique for symbolic execution in Python by overloading the bool method, allowing non-overloadable syntax to become overloadable. This approach enables recording paths through Python code and achieving symbolic execution without extensive modifications, particularly useful for integrating with Z3 solver and metaprogramming.

An AI-powered changelog generator with screenshots and customer reactions 

GitClear's changelog generator helps devs ditch the busywork of scouring git history to recall what was implemented. Snap Changelogs go on your website, GitHub profile, or anywhere a PNG can fit. SPONSOR

Fake Debugging II: The Race Condition Strikes Back

The post recounts debugging a complex multi-threading bug in a Python web framework, initially misdiagnosed as a shared mutable state issue. The author later discovers the true cause: an engine object inadvertently attached to a global class definition instead of a local namespace, leading to unexpected data leakage between requests in a multi-threaded environment.

Complex Django filters with Subquery

An overview on how to do complex filtering in Django with Subquery.

Symbolic Reference and Hardware Models in Python

The article discusses using symbolic models in Python for hardware design verification, focusing on an image downscaler example. It compares traditional hardware design flows with symbolic modeling, demonstrating how the latter can simplify debugging and verification by tracking data origins throughout the design process.

Project Quarantine

PyPI introduced a "Project Quarantine" feature to combat malware, allowing administrators to mark potentially harmful projects as quarantined, preventing easy installation while preserving project data. The post discusses the implementation details, usage statistics, and future plans for automation, aiming to reduce the time malicious projects remain accessible to users.

LAPACK in your web browser: high-performance linear algebra with stdlib

Implementing LAPACK routines for numerical computation in web applications.

Cowboy Debugging Python Memory Leaks

The article discusses how to use the objgraph library to debug memory leaks in Python. It also provides a number of tips and tricks for debugging memory leaks, such as using signal handlers to trigger ad-hoc behavior in a program and using a separate script to manage the state of the debugging session.

How to Code with AI in 2025 (For Non-Coders)

The video demonstrates how beginners can use AI to code in Python, covering setup and three projects: data scraping, bulk image editing, and creating a data dashboard.

Leadership Power Tools: SQL and Statistics

This is a post about leadership power tools. It discusses SQL and statistics. In order to make better decisions, leaders should be comfortable with extracting and interpreting data from stores. SQL is a powerful tool that can be used to query data. Statistical methods can be used to interpret data and make better decisions.

Interesting Projects, Tools, and Libraries

nv-ingest

NVIDIA Ingest is an early access set of microservices for parsing hundreds of thousands of complex, messy unstructured PDFs and other enterprise documents into metadata and text to embed into retrieval systems.

smolagents

A barebones library for agents. Agents write python code to call tools and orchestrate other agents.

Tuitorial

Create beautiful terminal-based code tutorials with syntax highlighting and interactive navigation.

IPychat

An AI extension for IPython to make it work like Cursor.

Termite

Generative UI in your terminal.

uv-migrator

A tool for migrating to the Python uv project.

magnetron

Minimalistic homemade PyTorch alternative, written in C99 and Python.

TangoFlux

Super Fast and Faithful Text to Audio Generation with Flow Matching.

LatentSync

Taming Stable Diffusion for Lip Sync!

KenobiDB

KenobiDB is a document-based data store abstraction built on Python’s sqlite3, offering a simple and efficient way to manage JSON-like data.

SiaPy

A tool for efficient processing of spectral images with Python.

Upcoming Events and Webinars

Virtual: PyLadies Berlin Meetup January 2025

There will be following talks

PuPPy Meetup January 2025

There will be a talk, You Should Build a Robot (MicroPython & Microprocessors).

PyData Berlin Meetup January 2025 

There will be following talks

Hybrid: PyData Chicago Meetup January 2025

There will be a talk, IOWarp: Advanced Data Management for Scientific Workflows.

Programmer Weekly - A free weekly newsletter for programmers.

Founder Weekly - A free weekly newsletter for entrepreneurs featuring best curated content, must read articles, how to guides, tips and tricks, resources, events and more.