 Hey, Bremen, quick question. Yeah, go for it. The Hangout on Air, is that accessible to any person with a Google account? Yes. OK, not just Wikimedia. No. Should I take, do you want to take it off the Media Wiki page, or? No, no, I think it's good. It's just there's somebody who has an if this, then add a thing in the list. OK, yeah, they'll be able to join. OK, see if they managed to hop on. Well, I think we should get started. Do you want to kick it off, check the broadcast, and everything? Yeah, I'll go ahead and start to stream now. And it's OK, we're live. Sorry about the siren in the background. I think it's siren testing time, where I live. Yeah, welcome to credit. It looks like we have at least a couple of demos this month. The first one is in the New Readers area, and Ann Gomez will be doing the demo, I believe. Is that right, Ann? Yes. OK. All right, so this is the first time I'm presenting a credit. So let me know if I do anything wrong here, and I can fix it. But basically, I wanted to share with you guys some prototypes that I've been working on, along with members of the reading team, to address some findings from the New Readers research that we did earlier this year. So specifically, we're working on trying to reach people who, or allow people to take content from Wikipedia from online to offline. So these are people who have internet access sometimes, or maybe are conscious of their data usage, or maybe they live in a city where, like New York even, where they ride the subway a lot. And it's useful to be able to have information available offline, but they're not really ready to go to our apps. So we're looking at a number of different ways to kind of meet this need. We'll be doing sort of deep user testing later this quarter. So the first one is, these are for mobile. I'm going to resize here. The first one is called WikiLater. These are working names, obviously. And basically, what this is, is an option for people to just download the PDF of the article in the full text. And we're currently using the lab's instance of the service. It's a little bit slow, so this will just take a second. But the cool thing about this is that it's all going to be optimized for mobile screens. So this will create the PDF to the size of the screen that you are using. This is a very short article, but you get the idea. The second one we're testing is called Flashcard for now. And I'm just going to pick an article. It's a little bit longer. So you can see this article has a lot of different sections. There's pictures. There's all sorts of information. And we're giving people the option to get just a summary of the article. And I think we're going to try to do this by a JPEG. We're figuring that out still. But that would be really potentially useful because it's a smaller file size, and it's a little bit more sort of intuitively shareable. So we've seen a lot of behaviors around people who do a lot of sharing on social media or sharing on WhatsApp groups with their friends or family. Interesting things, memes, quotes over pictures, all that kind of stuff. So we want to try to enable that. We still have some styling issues that we're figuring out here. But the idea would be to display just a picture, the title of the article, and the initial sort of summary. So a first paragraph and content from the info box. This looks like it's not rendering perfectly just yet. But we're still working on it. And then the third thing that we're testing is a progressive web app. This one's a little bit harder to demo on stream because you have to go offline to really show what it does. But what's cool about this is that it's all in the browser. So basically, we're letting people save articles to a reading list type thing. We're calling it saved pages for now. So they can look at their saved pages. And all of these are available offline. I would show you that. But if I turn off my internet, then I won't be able to stream anymore. Yeah, you can check these out. It's really cool. And then the other really neat thing about this is that we can prompt people, and we will prompt people to add this page to their home screen. So essentially what they'll have is an app-like experience on their device that allows them to have a shortcut to open up all of their saved articles that are available offline. And that's kind of the deal. Cool. Thanks, Ann. And it looks like Alangi. I'm not sure if I got the pronunciation correct. Has just joined. Alangi, can you hear me OK? Yeah, yeah. I hear you right. Cool. Yeah, are you ready to talk about automated testing and integration of if this, then that, support? Yeah, yeah, I'm ready. I think I'm ready. OK. OK, can I start? Yeah, are you sharing your screen? Yeah, let me just do that. It's that, oh, there you go. You got it. So can you see the screen? Yep, yep, we can see it. OK, so this project was done during the Google Summer of course, this 2016 session. And I had mentors, Stephen Laporte in San Francisco, Lisa Pinscher in Germany, Benedict Sedul in Germany, and Marius Hock in Germany. So the Wikidata team and Wikimedia US. So Stephen Laporte built a web service that connected the Wikipedia API and the EFT API. So in order to update users through EFT triggers, through a particular Wikipedia channel that was created on EFT, for users to have information like article of the day, picture of the day, and all that. So my project was divided into various parts to write automated tests for the triggers that were written by Stephen Laporte to build new triggers to support events on Wikidata and write the automated test and also to extend the application in such a way that it can support RSS feeds so that users will not need to connect to eft.com in order to get information or get updates about what is happening, the recent happenings. Whatever changes on Wikidata can be gotten through their feed-to-feeders. So I will just log in to my two-labs account. And also, this is actually the front end for the RSS feed. And OK, this is my two-labs account, and I'm going to use that tool. So this is the actual application running on two-labs, and it's using a Kubernetes web service to run. We hosted the application on Grid Engine, but after UVPanda actually deployed Kubernetes web service on two-labs, my primary mentor that was Stephen Laporte, he told me to change the web service that the application run on. So we are actually using Kubernetes right now. And I'll just activate my VM, and I'm going to run the automated tests that I wrote. While the test is running, I'll just run and open my, this is actually the codes for the testing part of the application. So the way it was written, I wrote a generic test that actually gets arguments, the triggers, and the ingredients, and create a test instance and run the test against it and get the results. If the test passes, then it returns an OK, and then if it fails, it returns a failure, and what happens? And this is the RSS. Meanwhile, there is a lot. The test will run very soon in a few seconds and print out the results. This is a UI for the RSS feeds. Let's take the example of the day, for instance. This is the RSS that is returned. So one part of the project with this link, you can subscribe to the RSS feed. Oh, I already have an article of the day. Let me delete this one and do a new subscription with that link. So there we go. These are the various articles of the day. So this is the latest that I received by email as a trigger on EVE this morning. So this is actually the article of today. It was sent, it was shoot out at 1 PM, UTC, I don't know. And these are the test results. So all of them passed. And when I was building this application, I had to make sure that since I'm using different APIs to connect the data, I had to make sure whatever I'm doing on EVE works on tool apps. And whatever I'm doing on tool apps works on EVE. So if the unit tests are running on tool apps, they should also, if it has, just give me a sec. Hello, is everybody getting me? Yeah, we see your EVE website. Yeah, I'm trying to access the developer site. And I'm redirected to partners. I really don't know what's going on here now. But the basic idea is whatever I do on tool apps for this test that run, I have to make sure that the same tests run on EVE. So I need to run an endpoint test against the application and make sure everything works. So it's a two-way thing. But the obvious thing is, if it passes here, it passes on EVE. And if it passes on EVE, it passes here. So this one makes sure that it's running on tool apps. Everything is OK. Maybe there is no box somewhere and all that. And if it also runs and then the channel passes. And for the RSS, subscribe and you get your RSS feeds. These are them. And basically, that is the project. This interface was just changed some few days ago. This, you had a different interface, but it was not looking very good. So I had to change it. And I'm still working on it in order to have the trigger descriptions and any other relevant information that would be useful for users to understand what the trigger is all about. So I'm improving on the API, the front end. And I'm also building new triggers. I'm also building new triggers to support wiki data. OK. Cool. So I think it'd be really interesting to see how your progress comes along in the future and the stuff. Thanks for the demo. You're welcome. Have a nice day. OK, you too. I think Stas, you're up next. OK, I'll do the screen sharing. So can you see the screen now? Yes. OK, so I'll be talking about two things today. The new file properties search and the new query service visualizations. So the first thing is that we have added the search by file properties or any media properties in wiki. It doesn't work yet on all wikis. Mainly, it doesn't work on comments because we need to do some reindexing, but it does work in this wiki. So for example, if we want to look for PDF files, we can look for files. And we can also look, for example, by size. So if we want large PDF files, we can look. So you see that the file size is larger than a megabyte and so on. So we have a bunch of things that we can ask for file, like type, size, resolution, bit depth, and so on, and all kinds of properties. So look at the documentation for the full list. But this allows you to search, basically, and limit your search to specific files. Of course, main interest would be on comments. So we will be enabling it on comments soon, too. And the second thing is visualizations for query service. So if you look at, for example, if you look at the query visualizations, we have added a number of new visualizations, like line chart, bar charts, cut-off chart, and area chart, that you can. OK, this is probably not a good example. Yeah, but you can try them out on different examples. And also, we have added one very interesting thing. You can build your own graph. So if you click on a graph builder, there is an interface that you have the columns from the query. And you can create, basically, your own graph. You can change all the properties. And then you click on Export. So you get this bunch of text. So what you do, you just go on the wiki. You just create the page, or use existing page. And you put a graph tag. And you just copy-paste whatever the graph builder produced. And you got a graph on wiki. And this graph is completely data-driven. So if you actually look in the source, there is a query inside. So every time the data on wiki data is updated, your graph is automatically updated. And you, of course, can change colors and stuff like that in the source. But that enables you much easier to