Technologie - 🤖

Bonjour à tous,

Depuis la dernière mise à jour de FreshRss, il est possible de faire du scrapping de site web avec xpath afin de créer des flux rss pour les sites qui n'en n'ont pas.

Plus de détail ici : https://danq.me/2022/09/27/freshrss-xpath/

J'ai mis en place ce principe pour récupérer des infos sur le site de la mairie et celui de la com'com.

Je bloque sur un troisième, le site de l'office de tourisme (pour récupérer l'agenda culturel). J'aurais besoin d'aide afin de m'en dépatouiller. Je ne sais pas si je suis au bon endroit. Mais si vous savez vers quelle communauté je pourrais me tourner pour résoudre ce soucis, n'hésitez pas à me le dire.

Un peu de contexte.

Lors de son chargement, le site de l'office de tourisme charge dans un premier temps le html, mais sans les événements de l'agenda. Ces événements sont présents sur la page, mais dans une balise script sous forme d'un json stocké dans une variable item.

Une autre requête POST est ensuite effectuée avec en paramètre ce json pour le mettre en forme (html) puis l'insérer dans le corps de la page.

À partir de là, le scrapping se complexifie, auriez-vous une idée de comment récupérer ce json pour le transformer en quelque chose qui ressemblerais à un flux atom/rss/json ?

Je pensais faire un script bash, mais je suis un peu perdu avec xmlint, xisdel et jq.

C'est peu être un peu ambitieux, qu'en pensez-vous ?

cross-posted from: https://jlai.lu/post/4037671

Links to the live streaming in each room will appear on this page during the conference.

Static overviews of the rooms and the room occupation by track are available, too.

FOSDEM is a two-day event organised by volunteers to promote the widespread use of free and open source software.

Usually taking place in the beautiful city of Brussels (Belgium), FOSDEM is widely recognised as the best such conference in Europe

What is FOSDEM? 

FOSDEM is a free and non-commercial event organised by the community for the community. The goal is to provide free and open source software developers and communities a place to meet to:

get in touch with other developers and projects;

be informed about the latest developments in the free software world;

be informed about the latest developments in the open source world;

attend interesting talks and presentations on various topics by project leaders and committers;

to promote the development and benefits of free software and open source solutions.

Participation and attendance is totally free, though the organisers gratefully accept donations and sponsorship.

cross-posted from: https://lemmy.ml/post/11232276

The idea comes from the discussion that occurred over the new lemmy-clone or alternative, SubLinks and how its main feature is that its tech stack is different from lemmy's which should enable all of the developers who don't know rust to contribute.

One of the core lemmy devs (dessalines I believe) said responded to these general sentiments by saying something to the effect of rust being a good technical choice and that learning rust in order to contribute would be a good expenditure of time (as tech people need to learn new things all the time anyway).

Soooo ... for those interested ... how about we all learn together rust through learning about ActivityPub and Lemmy's codebase and solving problems and making contributions? We could have a community dedicated to asking questions, sharing solutions or ideas and generally discussing all things we're learning about rust, activitypub, fediverse and lemmy? If an actual community can be built around the desire to learn rust and give back to lemmy with all us newbs working together as much as posslbe ... that would have to be a win right?

Even better if those who know more about the topic could use the community as a chance to post or write up what they know for us to learn from. For instance, I've glanced at lemmy's code base (without knowing rust of course) and I feel like it could do with an architecture birds-eye perspective on how the code base works.

Obviously chatting on matrix might be a good place for this, especially as devs and admins are chatting there already ... but I feel like the structure of lemmy might be a better place for a sort of reading club.

Any thoughts or takers? I feel like creating the community on lemmy.ml would make sense, maybe having one of the core devs as a mod too?

cross-posted from: https://lemmy.ca/post/14221826

cross-posted from: https://lemmy.dbzer0.com/post/13215898

cross-posted from: https://lemmy.dbzer0.com/post/13211160

Simply mention @aihorde@lemmy.dbzer0.com with the text "draw for me " and then the text that you would like it to draw and I will draw you an image matching your prompt, post it in this community and then reply with a link to each image! Optionally you can provide a "style:" as well. Example: [@aihorde@lemmy.dbzer0.com](https://lemmy.dbzer0.com/u/aihorde) draw for me a beautiful night style: fantasy

This is a free service provided through the AI Horde. Please consider supporting us on Patreon or onboarding your own worker onto the horde if you enjoy this bot!

Source Code

cross-posted from: https://lemmy.ml/post/11023519

This is a chance for any users, admins, or developers to ask anything they'd like to myself, @nutomic@lemmy.ml , SleeplessOne , or @phiresky@lemmy.world about Lemmy, its future, and wider issues about the social media landscape today.

NLNet Funding

First of all some good news: We are currently applying for new funding from NLnet and have reached the second round. If it gets approved then @phiresky@lemmy.world and SleeplessOne will work on the paid milestones, while @dessalines and @nutomic will keep being funded by direct user donations. This will increase the number of paid Lemmy developers to four and allow for faster development.

You can see a preliminary draft for the milestones. This can give you a general idea what the development priorities will be over the next year or so. However the exact details will almost certainly change until the application process is finalized.

Development Update

@ismailkarsli added a community statistic for number of local subscribers.

@jmcharter added a view for denied Registration Applications.

@dullbananas made various improvements to database code, like batching insertions for better performance, SQL comments and support for backwards pagination.

@SleeplessOne1917 made a change that besides admins also allows community moderators to see who voted on posts. Additionally he made improvements to the 2FA modal and made it more obvious when a community is locked.

@nutomic completed the implementation of local only communities, which don't federate and can only be seen by authenticated users. Additionally he finished the image proxy feature, which user IPs being exposed to external servers via embedded images. Admin purges of content are now federated. He also made a change which reduces the problem of instances being marked as dead.

@dessalines has been adding moderation abilities to Jerboa, including bans, locks, removes, featured posts, and vote viewing.

In other news there will soon be a security audit of the Lemmy federation code, thanks to Radically Open Security and NLnet.

Support development

@dessalines and @nutomic are working full-time on Lemmy to integrate community contributions, fix bugs, optimize performance and much more. This work is funded exclusively through donations.

If you like using Lemmy, and want to make sure that we will always be available to work full time building it, consider donating to support its development. Recurring donations are ideal because they allow for long-term planning. But also one-time donations of any amount help us.

• Liberapay (preferred option)
• Open Collective
• Patreon
• Cryptocurrency

salut les gens.

J'ai plusieurs pc portables KO et je voudrai transformer les écrans de ceux-ci en écran externe.

Le seul problème est que je ne suis pas sure du nom du connecteur qui connecte l’écran du pc portable à sa carte mère.

Dans plusieurs vieux articles de blog / messages de forum, je trouve mention d'une connectique LVDS. Cependant j'ai des doutes en voyant certaines des images trouvée avec la recherche "lvds connector" via DuckDuckGo

Sachant que ce qui y ressemble le plus serait plutôt "laptop lcd cable replacement"

1. "LVDS": est il bien le nom du connecteur ?
2. Savez vous si il existe des convertisseurs vers du HDMI/DisplayPort ?
3. Si quelqu'un a déjà tenté l'aventure et a des conseils je suis preneur.

Merci

publication croisée depuis : https://lemmy.world/post/11075122

Émission de Xavier de la Porte, autour de la thèse de sociologie de Camille Girard-Chanudet.

Passionnante plongée dans le quotidien d'une équipe d'annotatrices pour entrainer un algorithme d'anonymisation de décisions de justice :

On perçoit à quel point les catégories d'annotation pèsent lourd, et combien il est crucial que l'annotation humaine se place dans une perspective d'intelligence collective.

Derrière un imaginaire peuplé d’androïdes rutilantes se cachent les petites mains qui annotent patiemment les données. Un travail sans fin, car le réel sans cesse changeant échappe à l'automatisation.

La thèse de sociologie de Camille Girard-Chanudet éclaire les rouages du deep learning : au palais de justices, des travailleuses de catégorie C travaillent avec application à annoter les données pour un algorithme d'anonymisation des décisions de justice destinées (entre autres) aux startups des legal techs.

L'émission de Xavier de La Porte révèle l'ampleur des collaborations humaines encapsulées dans les machines que l'on voudrait nous présenter comme intelligentes. C'est une myriade de décisions qui conduisent à définir des catégories réductrices pour saisir le réel. Ces catégories s'imposent silencieusement - d'abord aux travailleuses et travailleurs du clics, puis aux utilisatrices et utilisateurs des services d'IA.

Toutefois, à la différence de la majeure partie du "digital labor" étudié par Antonio A. Casilli et ses collègues du programme Panoptiwork, cet exemple démontre qu'il est possible de mobiliser l'IA dans une perspective d'intelligence collective.

L'intelligence collective, c'est "la mise en commun de la réflexivité afin de finaliser une action qui ne pourrait pas l'être par une seule personne" (Samuel Szoniecky et Nasreddine Bouhaï). Cette mise en commun n'est possible qu'au prix de contraintes destinées à rendre le travail interopérable. Ainsi, l'activité des annotatrices du Palais de Justice est cadrée par des catégories, mais leur position pérenne leur permet d'interroger et de faire évoluer ces catégories.

Les grands modèles de NLP sur lesquels s'appuient de telles applications restent ceux que des entreprises privées daignent partager en open source. Ils ne sont pas entrainés dans des conditions aussi idylliques, loin de là. Est-ce tolérable ?

Issu d'un com de LinuxFR :
https://github.com/cnlohr/mini-rv32ima > Emulateur RISC-V minimal 😍

Au service du capital la tech, ofc.
Le libre, cest politique avant tout.

cross-posted from: https://lemmy.dbzer0.com/post/12677754

cross-posted from: https://lemmy.dbzer0.com/post/12677747

cross-posted from: https://beehaw.org/post/11177244

"All the bugs and errors have been known at one level or not, for many, many years. Right from the very start of deployment of the system, there were bugs and errors and defects, which were well-known to all parties," said Paul Patterson, co-CEO of Fujitsu's European division. Ars Trending Video

That goes back to 1999, when the Horizon software system was installed in post offices by Fujitsu subsidiary International Computers Limited. From 1999 to 2015, Fujitsu's faulty accounting software aided in the prosecution and conviction of more than 900 sub-postmasters and postmistresses who were accused of theft or fraud when the software wrongly made it appear that money was missing from their branches.

On lui doit le protocole réseau NTP qui permet de mettre à l'heure les ordinateurs sur Internet.

https://mastodon.social/@univbordeaux pour le compte Mastodon