1140
News & links about Python programming. https://pythonhub.dev/ Administrator: @rukeba
Timesliced reservoir sampling: a new(?) algorithm for profilers
Reservoir sampling lets you pick a sample from an unlimited stream of events; learn how it works, and a new variant useful for profilers.
https://pythonspeed.com/articles/reservoir-sampling-profilers/
From zero to a RAG system: successes and failures
Building a production RAG system is far more about data pipelines, indexing strategy, and infrastructure tradeoffs than model choice, with most failures coming from scaling, retrieval quality, and compute constraints. The key lesson is that RAG success depends on iterative engineering and system design discipline, not just plugging in an LLM, with real-world performance shaped by bottlen...
https://en.andros.dev/blog/aa31d744/from-zero-to-a-rag-system-successes-and-failures/
Autograd and Mutation
How does PyTorch autograd deal with mutation? In particular, what happens when a mutation occurs on a view, which aliases with some other tensor? In 2017, Sam Gross implemented support for in-place operations on views, but the details of which have never been described in plain English… until now.
https://blog.ezyang.com/2026/03/autograd-and-mutation/
Building a Navier-Stokes Solver in Python from Scratch: Simulating Airflow
A hands-on guide to implementing CFD with NumPy, from discretization to airflow simulation around a bird's wing
https://towardsdatascience.com/building-a-navier-stokes-solver-in-python-from-scratch-simulating-airflow/
NumPy as Synth Engine
NumPy can be used as a real time sound synthesis engine, generating all audio directly from mathematical functions like waves, noise, and filters without any pre recorded samples. The broader idea is that powerful general purpose tools like NumPy can be pushed far beyond their intended use, enabling complex systems like music generation through pure computation.
https://kennethreitz.org/essays/2026-03-29-numpy_as_synth_engine
How Clean Code Turns Into Overengineering
This video is about how code that looks clean can still hide a bad design, and why overusing tiny abstractions can make a program harder to understand and change. It refactors a Python reporting example by simplifying the structure, making the pipeline explicit, and focusing on cohesion over smallness.
https://www.youtube.com/watch?v=U4sPMwAiXco
diffly
Utility package for comparing polars data frames.
https://github.com/Quantco/diffly
Fixed Python autocomplete
The post suggests that heavy LSP and static analysis approaches are unnecessary for many common autocomplete scenarios. It shows a lightweight, pattern-based approach can deliver faster, more responsive suggestions without full semantic analysis.
https://matan-h.com/better-python-autocomplete
Pydantic AI - Intro to Agentic AI with Pydantic AI framework
We'll look at using Pydantic AI to build agent-based workflows, starting with simple fundamentals, and building up to more complex examples that use vector databases, RAG, multi-agent workflows and more.
https://www.youtube.com/playlist?list=PL-2EBeDYMIbSWGoDzOFm33_5W_ShO-VIi
Python Hub Weekly Digest for 2026-04-05
https://pythonhub.dev/digest/2026-04-05/
rsloop
An event loop for asyncio written in Rust.
https://github.com/RustedBytes/rsloop
The Hidden Mechanism Behind Clean Python APIs (Descriptor Deep Dive)
Descriptors define how Python resolves attribute access, explaining why values sometimes come from the instance, class, or elsewhere in non-obvious ways. Understanding descriptor rules enables cleaner, more reusable designs by giving you precise control over attribute behavior.
https://www.youtube.com/watch?v=7SUzTOkUVLY
Visitran
Build data transformation pipelines using Python with a visual IDE and AI assistant.
https://github.com/Zipstack/visitran
Would it have been better if Meta bought Astral.sh instead?
https://www.reddit.com/r/Python/comments/1ryglss/would_it_have_been_better_if_meta_bought_astralsh/
OpenSpace
Make Your Agents: Smarter, Low-Cost, Self-Evolving.
https://github.com/HKUDS/OpenSpace
ATLAS
Adaptive Test-time Learning and Autonomous Specialization.
https://github.com/itigges22/ATLAS
django-modern-rest
Modern REST framework for Django with types and async support!
https://github.com/wemake-services/django-modern-rest
vectorize-io / hindsight
Hindsight: Agent Memory That Learns
https://github.com/vectorize-io/hindsight
Why pylock.toml includes digital attestations
Including digital attestations in pylock.toml allows developers to verify the origin and integrity of dependencies, not just their versions and hashes, improving protection against supply chain attacks. The broader point is that modern package security requires provenance, not just reproducibility, so lock files are evolving from “what to install” into “what can be trusted to install.”
https://snarky.ca/why-pylock-toml-includes-digital-attestations/
claude-howto
A visual, example-driven guide to Claude Code - from basic concepts to advanced agents, with copy-paste templates that bring immediate value.
https://github.com/luongnv89/claude-howto
Smello
A developer tool that captures outgoing HTTP requests from your code and displays them in a local web dashboard.
https://github.com/smelloscope/smello
Oxyde ORM
A type-safe, Pydantic-centric asynchronous ORM with a high-performance Rust core designed for clarity, speed, and reliability.
https://github.com/mr-fatalyst/oxyde
agentscope-ai / ReMe
ReMe: Memory Management Kit for Agents - Remember Me, Refine Me.
https://github.com/agentscope-ai/ReMe
justx
A TUI command launcher built on top of just. Define recipes once, run them anywhere.
https://github.com/fpgmaas/justx
Build Your Own Openclaw - A step by step guide, using python
https://github.com/czl9707/build-your-own-openclaw
LiteLLM Python package compromised by supply-chain attack
https://github.com/BerriAI/litellm/issues/24512
OpenAI to acquire Astral
https://www.reddit.com/r/Python/comments/1rxzy4d/openai_to_acquire_astral/
Reinventing Python's AsyncIO
The post explores a redesign of Python’s async runtime, arguing that the current async/await and event-loop model adds unnecessary complexity, and proposing a simpler runtime where concurrency is handled automatically without explicit async syntax.The author experiments with a new runtime approach that can run async workloads 2–3.5× faster than traditional asyncio, suggesting Python’s co...
https://blog.baro.dev/p/reinventing-pythons-asyncio
How we optimized Dash's relevance judge with DSPy
Dropbox used DSPy to turn prompt engineering for our relevance judge into a measurable, automated optimization loop, improving task performance, cost, and how reliably it works in production.
https://dropbox.tech/machine-learning/optimizing-dropbox-dash-relevance-judge-with-dspy