PythonHub

onyx-dot-app / onyx Open Source AI Platform - AI Chat with advanced features that works with every LLM https://github.com/onyx-dot-app/onyx

Python Hub Weekly Digest for 2025-10-05 https://pythonhub.dev/digest/2025-10-05/

Air The new web framework that breathes fresh air into Python web development. Built with FastAPI, Starlette, and Pydantic. https://github.com/feldroy/air

Unlocking Performance in Python's Free-Threaded Future: GC Optimizations A description of the performance optimizations made to the free-threaded garbage collector for Python 3.14. https://labs.quansight.org/blog/free-threaded-gc-3-14

LLMs from Scratch – Practical Engineering from Base Model to PPO RLHF This video provides a hands-on guide to building a large language model entirely from scratch in PyTorch, covering every step from core transformer design to advanced alignment with RLHF. By the end, viewers gain practical experience in implementing, training, scaling, and aligning their own custom LLMs. https://www.youtube.com/watch?v=p3sij8QzONQ

Python Singleton Pattern: Smarter Than You Think? This video analyzes the strengths and weaknesses of the singleton pattern in Python, explaining why global state is risky but controlled instantiation can be valuable in certain cases. It recommends module-level singletons and thread safety measures, while cautioning against tight coupling and testing pitfalls with traditional singleton implementations. https://www.youtube.com/watch?v=p_UQ7tzUFLo

How to Build Advanced AI Agents – Course for Beginners (LiveKit, Exa, LangChain) The video teaches beginners how to build advanced AI agents, such as voice sales agents, research assistants, and multi-agent workflows, using LiveKit, Exa, LangChain, and Cerebras. It provides step-by-step guidance, hands-on code, and free API credits to help developers quickly create real-world AI applications. https://www.youtube.com/watch?v=B0TJC4lmzEM

The Kaggle Grandmasters Playbook: 7 Battle-Tested Modeling Techniques for Tabular Data The Kaggle Grandmasters Playbook presents seven proven techniques for tabular data modeling, emphasizing fast experimentation and careful validation powered by GPU acceleration to handle large-scale data effectively. Key strategies include advanced exploratory data analysis, building diverse baselines, extensive feature engineering, ensembling with hill climbing and stacking, pseudo-labe... https://developer.nvidia.com/blog/the-kaggle-grandmasters-playbook-7-battle-tested-modeling-techniques-for-tabular-data/

LLM-Deflate: Extracting LLMs Into Datasets LLM-Deflate is a technique for systematically extracting structured datasets from trained large language models by probing their internal knowledge with hierarchical topic exploration and prompt engineering. This reverse-compression process enables model analysis, knowledge transfer, training data augmentation, and debugging, potentially making knowledge extraction a standard tool as inf... https://www.scalarlm.com/blog/llm-deflate-extracting-llms-into-datasets

Cloud-Native Pipelines for Scientific Data Processing with Prefect and Dask This article explains how to build scalable, cloud-native scientific data processing pipelines using Prefect for workflow orchestration and Dask for parallel computation. It covers cloud-optimized formats (like Zarr), integration with tools like xarray and echopype, and demonstrates end-to-end ETL pipelines that load, process, and store multidimensional data directly in the cloud. https://oceanstream.io/cloud-native-data-processing-pipelines-with-prefect-and-dask/

How Well Do New Python Type Checkers Conform? A Deep Dive into Ty, Pyrefly, and Zuban The Python type checking landscape in 2025 includes three new Rust-based tools: Astral's ty, Meta's pyrefly, and Zuban. Ty emphasizes gradual adoption with fewer false positives, pyrefly focuses on aggressive inference to catch more issues early, and Zuban aims for seamless mypy compatibility; while conformance tests reveal differences, all show promise for real-world Python development. https://sinon.github.io/future-python-type-checkers/

OM1 Modular AI runtime for robots. https://github.com/OpenMind/OM1

enso enso is a functional programming framework for Python. https://gitlab.com/evansemenoff/enso

MapAnything Universal Feed-Forward Metric 3D Reconstruction https://github.com/facebookresearch/map-anything

Compiling Python to Run Anywhere The article discusses an innovative approach to compiling Python code into cross-platform, ahead-of-time optimized machine code executables without modifying the original Python source. It details building a custom symbolic tracer, propagating types for lowering to C++, leveraging AI to generate C++ operators, and empirically optimizing performance across multiple hardware targets to ena... https://blog.codingconfessions.com/p/compiling-python-to-run-anywhere

fastapi-radar A powerful debugging dashboard for FastAPI applications. Monitor HTTP requests, SQL queries, and exceptions in real-time with a beautiful React UI. One-line integration, zero configuration needed. https://github.com/doganarif/fastapi-radar

VectorLiteDB The SQLite for vector embeddings — A simple, embedded vector database that stores everything in a single file. https://github.com/vectorlitedb/vectorlitedb

How I used Cursor AI to migrate a Bash test suite to Python The migration of a large Bash container test suite to Python using the Cursor AI code editor saved about 1.5 months of development time, with Cursor handling script conversion, function replacement, and automated PyTest suite generation. Although the migration was not entirely smooth and required some manual fixes, the resulting Python test suite passed tests successfully, demonstrating ... https://developers.redhat.com/articles/2025/09/23/how-i-used-cursor-ai-migrate-bash-test-suite-python#our_real_world_results

Python-Style Kwargs in TypeScript https://xavd.id/blog/post/python-kwargs-in-typescript/

Klavis MCP integration layers that let AI agents use thousands of tools reliably. https://github.com/Klavis-AI/klavis