Juillet-Août 2023

Bonjour, voici les nouvelles de la rentrée. N’oubliez pas le café Julia le 15 septembre à 10:00 et les Journées Julia et Optimisation du 4 au 6 octobre au CNAM Paris.

La France passe troisième au nombre d’utilisateurs derrière l’Allemagne et les Etats-Unis d’après les résultats du sondage annuel réalisé pour JuliaCon.

Julia est entré dans le top 20 du TIOBE Index qui classe les langages de programmation par popularité.

Il y a maintenant une centaine de nouveaux packages par mois donc vous ne voyez ici qu’une sélection personnelle. N’hésitez-pas à m’envoyer vos liens ou vous pouvez faire directement une PR sur https://github.com/pnavaro/NouvellesJulia.

Une information importante a été publiée cet été à propos de l’utilisation des threads. Je conseille vivement de lire cet article:

PSA: Thread-local state is no longer recommended

Billets de blog¶

• Bogumił Kamiński - How does Tables.jl handle schema-less tables?

• Ronny Bergmann - Render Quarto Tutorials in Documenter.jl with GitHub Actions

• Frames White - JuliaLang Antipatterns

• Cameron Pfiffer - What the heck is Julia and why won’t Cameron shut up about it

• Matthijs Cox - User-defined Show Method in Julia

• Alexandra Gessner - Supervised learning: One step at a time

• Panagiotis Georgakopoulos - JuliaHub: Ask AI

Packages¶

• AbstractNeuralNetworks.jl : Abstract data structures for the construction of neural networks.

• ArviZ.jl : Exploratory analysis of Bayesian models.

• AutoVectors.jl : Julia vectors whose first and last indices are any integers.

• BioMarkovChains.jl : Representing biological sequences as Markov chains.

• ChebyshevFiltering.jl : A package to perform large-scale sparse diagonalization.

• CounterfactualExplanations.jl : A package for Counterfactual Explanations and Algorithmic Recourse in Julia.

• CurveFitParameters.jl : Curve fitting problems.

• DataFlowTasks.jl : Tasks which can keep track of how data flows through it.

• Deneb.jl : API for creating Vega-Lite visualizations.

• Diffractor.jl : Experimental next-generation, compiler-based AD system.

• DomainColoring.jl : Domain colorings and checker plots of complex functions using smooth colors.

• Eikonal.jl : Solvers for the Eikonal equation.

• FloatTracker.jl : Library providing tracking of floating point errors through a program resources.

• GaussianKDEs.jl : Gaussian Kernel Density Estimators.

• GeometricMachineLearning.jl : Structure Preserving Machine Learning Models.

• GradValley.jl : Lightweight package for Deep Learning.

• HeterogeneousComputing.jl : Tools for heterogeneous computing.

• IncompressibleNavierStokes.jl : Incompressible Navier-Stokes solver.

• IPUToolkit.jl : Interface the Intelligence Processing Unit (IPU) by Graphcore.

• JuliaTDA : Topological Data Analysis organization.

• KinematicCoordinateTransformations.jl : Perform coordinate system transformations of position, velocity, acceleration, and jerk.

• LanguageModels.jl : Load nanoGPT-style transformers.

• LowRankMatrices.jl : Lightweight package that provides the low-rank matrix types that are used in LowRankApprox.jl.

• MakiePublication.jl : Package for producing publication quality figures based on Makie.jl.

• MoYe.jl : Layout Algebra on GPU.

• MultilayerGraphs.jl : A package for the creation, manipulation and analysis of the structure, dynamics and functions of multilayer graphs.

• MultivariateChebyshev.jl : Interpolation of functions using Chebyshev polynomials.

• MulticomplexNumbers.jl : Package for representing multicomplex numbers and performing multicomplex algebra.

• NormalizingFlows.jl : A normalizing flow library.

• NUMA.jl : Tools for querying and controlling NUMA policies.

• Parsers.jl : Fast parsing machinery for basic types.

• PlotlyDocumenter.jl : Show plotly plots in Documenter.jl as static HTML.

• ReferenceFiniteElements.jl : educational and lightweight package to define finite elements in the reference (quadrature space) configuration.

• SemanticAST.jl : Semantic analysis for Julia source code.

• SimpleChains.jl : Neural network for small problems on the CPU.

• SimplePolynomialRing.jl : Polynomial ring realization.

• SoleModels.jl : Symbolic modeling and learning.

• SVDSubset.jl : Fast and memory-efficient implementation of svds algorithm.

• TensorTrains.jl : Tensor Trains, mostly thought of as probability distributions.

• TidierPlots.jl : Julia implementation of the ggplot2 R package.

• TimeseriesTools.jl : A convenient package for working with time series as mathematical series, rather than date-indexed data structures.

• VoronoiFVM.jl : Solution of nonlinear multiphysics partial differential equation systems using the Voronoi finite volume method.

Vidéos¶

• Christopher Rackauckas - Fast Neural ODE / UDE: Improved Parallelism and Memory Performance Differentiating Stiff ODEs

• Abel Siqueira - Crash course on Julia Basics - Loop, function, types, Dict, indexing

• Julia Tuber - Top 10 JULIACON

• Rémi Vezy - For loops in Julia (eachindex, enumerate...)

• Jakob Nissen - Minimise Julia’s Latency: Revise, PrecompileTools, and PackageCompiler

• Jon Doucette - Ignite.jl: A Brighter Way to Train Neural Networks

• Kyle Daruwalla - Intro to FluxML and Machine Learning in Julia

• Francois-Xavier Coudert - CrystalNets: topology identification of crystalline materials

• James Schloss - Breaking up with Julia

Tutoriels, Documentation¶

• Adrian Salceanu et al - Genie Dashboard showing a real-time simulation of the Lorenz equations

• Yolhan Mannes - A tutorial to use Flux on the MNIST DataSet

• Patrick Altmeyer - Julia for Deep Learning

• Uwe Fechner - Working with Julia projects

• Guillaume Dalle et al - A series of blog posts on best practices for Julia development

• Carsten Bauer - How to provide Julia to users on HPC clusters ?

• Andrew Gelman - Regression and Other Stories

• Jakob Nybo Nissen - How to optimise Julia code: A practical guide

Pierre

Abonnez-vous à la liste Julia du CNRS

Archives