MIT licensed document extraction Python library from the Deep Search team at IBM, who released [Docling v2](https://ds4sd.github.io/docling/v2/#changes-in-docling-v2) on October 16th. Here's the [Docling Technical Report](https://arxiv.org/abs/2408.
How does the Markdown output from Docling compare with the HTML that you've gotten out of Gemini for PDF documents? Does Docling do a good job of recognizing headings, lists, etc.?
The most recent Jerry Rig Everything video is deep, profound, poignant and worth a watch. Americans, vote. Please, just vote. youtu.be/nZftVjs-JdY?si=9mqKyU…
Life is short - why make it harder? The USA is the *only* wealthy country in the world who does not have a form of Universal healthcare. Why get rid of what ...
The Five Faces of Doctor Who, the multi-Doctor repeat season was broadcast 40 years ago today, on 2 November 1981. This fondly remembered run of repeats gave many fans the chance to enjoy stories from past Doctor’s that they’d never seen before for the first time! #DoctorWho
@DavidGoldfield IKWYM, it still baffles me why they've never repeated the show on a regular basis, they seem to with just about everything else these days.
I am thinking of using github.com/keithamus/csslex as a base and going from there to learn more about CSS internals, but I admit I am Afraid of Doing Code Crimes and would like encouragment ha ha
not with that library in particular, but in librsvg I use a rust-based css lexer and parse properties with that. Writing the parsers is fun, a bit tricky sometimes. I do steal code from Servo from time to time. Happy to talk about it.
Oh the non believers among you re my toot about Woollybutt St!! I had to walk 5km there and back this . morning to bring you this .... It even has its own dedicated bus stop (@welshpixie)
He may be starting with lanes in Toronto, but all communities are at risk, including Ottawa. We need to stand up for everyone who uses these lanes, benefits from the cyclists shopping along the routes, and all those who will be too afraid to ride because of the increased danger. See you there! #ottbike #biketooter
Shall the historic anglophones ALSO sue the province of Québec AND the city of Montréal, etc. to translate ALL its websites and provincial and municipal documents into English?
MONTREAL — A Quebec civil rights group is suing the office of the registrar of the Supreme Court of Canada because of the high court’s refusal to translate its historic decisions into French.
To pull a Bezos: to sell oneself out to someone only to be kicked in the teeth by them.
Couldn’t have happened to a smarter person. Like literally, almost anyone paying attention with a modicum of intelligence would have known acquiescence would earn them nothing but scorn.
It was a difficult day for Montreal metro users on Nov. 1 with service only being restored in the early morning of Nov. 2.There were multiple delays including the green line being at a standstill for 20-hours.
Stone-built Snake Bridge on the Macclesfield Canal in England.
The ‘Snake Bridge’ or 'Roving Bridge' is a bridge over a canal constructed to allow a horse towing a boat to cross the canal when the towpath changes sides, without having to be unhitched from the boat.
Trump Media Outsourced Jobs to #Mexico Even as #Trump Pushes “America First” ==
On the campaign trail, #DonaldTrump threatened businesses that send jobs south of the border, while his own company that runs the Truth Social platform outsourced coding jobs to workers in Mexico, outraging some staff members.
On the campaign trail, Donald Trump threatened businesses that send jobs south of the border, while his own company that runs the Truth Social platform outsourced coding jobs to workers in Mexico, outraging some staff members.
If you buy a Mac with Sequoia on their closed arm SOC you're contributing to the enshittification of general purpose computing and will get the feudalized computing experience webserfs crave. hackaday.com/2024/11/01/apple-…
Many MacOS users are probably used by now to the annoyance that comes with unsigned applications, as they require a few extra steps to launch them. This feature is called Gatekeeper and checks for …
There's a lot more I could have written about this but... look we've all seen this a million times before, we know how it works, how bad actors abuse it to mass report and bully and where it all goes.
Hey folks, thanks to @bagder and @kees , I managed to compute the "Source code age" from the GLPI project, which I'm working on, and it's really fun because we can see a lot of nice information on it!
To sum it up: the code is very dynamic, and compared to Curl (which you can see here: mastodon.social/@bagder/113399… ) , we can definitely see that it doesn't endorse backwards compatibility: there are a lot of refactorings of big parts of the codebase! 🎉
I already sent it to the GLPI internals, and explained how to read the graph, I hope they will be as thrilled as I am! Gosh, I love statistics like that so much I could make it my entire job 😅
Yes, I did find them, it makes me want to create a repo that would generalize such graphics for any repository! Would be quite hard work, but at least adapting one-by-one and generating the stats is already nice!
Though I must admit that your scripts are quite slower, and they're really curl-specific. The python script by @kees for the Linux Kernel is a bit more generic and easier to standardize.
Would take a while to adapt all of these incredible graphics for all repositories, but maybe it's worth it 😅
I'm sure the scripts can be done better and faster and certainly more general purpose. My ambition has been to cover the curl project from as many angles as possible rather than to make generic scripts.You know, limited time and energy and all that.
I think you have some non-main-branch tags in your data set. E.g. The '22 code drops almost to 0 in '23 but then _comes back_ in '24? It shouldn't be possible for that to happen: once it's removed it can't come back. LoC for a given year span can only grow during that year and then it must only shrink.
Just like Symfony and some other frameworks, GLPI has to maintain multiple versions at a time. So I should actually display only the version tags for a specific branch (9.x for instance) and re-run the script alter with the tags from another branch (like 10.x).
The gethostbyname(3) family was removed in POSIX-1.2008 in favor of getaddrinfo(3) introduced in POSIX-1.2001. Modern POSIX systems such as Sortix does not have gethostbyname nor the related defini...
Táta půjde na operaci s kolenem a na dobu, co bude marodit, mi půjčil pracovní nebeskou Felicii, abych jí udržovala v provozu 💙 Jízda tichá, nic nebouchá, světla svítí, stěrače stírají, bez posilovače řízení, vepředu i vzadu švihadla, přidaný denní svícení a dálkový centrální zamykání... prostě stará dobrá klasika bez zbytečný elektroniky 😀👌🏻
Thirty six years almost to the day after the release of the Morris Worm, OKTA announces they've just patched a bug where you can just log in with no password if your username is too long.
Could we please, before the Morris Worm turns forty years old, make a habit of, god help us all maybe even a standard practice of, sanitizing our inputs.
Encryption backdoors must never be allowed. To prove that Tuta Mail and Tuta Calendar are free from any backdoor, the entire client code is published as open source. Let's fight against mass surveillance! ✊
A very good way to have instant access to all program knowledge, workings, creations, and language. The easy access blueprints to robbing your entirety and wiping you clean.
Then proceed with, "This is our version, new, updated from the ground up, a re-creation of a familiar past creation."
Intelligence officials say the video, which purported to show a Haitian immigrant claiming he had voted multiple times in Georgia, is the product of a Russian propaganda operation.
Die #Musik von #Silentlive hörte ich vor einigen Jahren in einem kleinen #Café in #Wałbrzych - ich versuchte der Kellnerin verständlich zu machen, dass ich die CD gern einfach sofort kaufen würde. Nach einigem hin und her verkaufte sie mir ihr eigenes Exemplar. Eine schöne Erinnerung.
In my 20 years of podcasting, I’ve tried many services for recording podcast interviews. I’ve found nothing better than Cleanfeed, because it’s reliable and flexible. While there are services that record locally and then upload, all of them glitch out at one time or another, because there is a lot of complexity in getting that right. It’s heartbreaking to lose a good interview because your double-ender service had a bad hair day. It happens far too much. Cleanfeed is also good for broadcasting and other live events. During the lockdowns, many broadcasters were using Cleanfeed to send entire shows, including high quality stereo music streams, back to radio stations. I did an explanation and demonstration of Cleanfeed in The Blind Podmaker podcast which is still relatively current. I say all this because Cleanfeed has just come out with a very nice update for iPhones using USB-C. You can now use a USB-C audio device including audio interfaces, come into Cleanfeed through safari for iOS, and Cleanfeed will recognise the audio device. This is going to be a big deal for people who want to participate in high quality podcasts or live broadcasts while on the go. If a radio station has a broadcaster in the field for some reason, coming in through Cleanfeed and a USB-C audio device is going to sound way better than a FaceTime call. Best of all, Cleanfeed’s developers care deeply about #accessibility. They’ve responded positively to many of my suggestions and those of others in the blind podcasting community. You can find Cleanfeed at cleanfeed.net.
I had a talk called "Unbreakable Linux for Entire Family" today. It had nothing to do with Oracle #Linux, it was about concepts and approaches around #Fedora #Silverblue to make a truly free ChromeOS-like system for friends and family. I was happy to have a full room of people. It's good to see #OpenAlt bouncing back in terms of attendance.
))
Matt Campbell
in reply to Simon Willison • • •Simon Willison
in reply to Matt Campbell • • •