[09:45:24] New patchset: Mwjames; "SMW\FeedResultPrinter to move into the SMW namespace" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47570 [10:01:01] Hello, anyone here to help me with automatic login in PHP code? [10:03:30] Mariana__: Is it related to SMW in any way? [10:08:55] Nischay: No, i opened mediawiki channel, thanks! [10:09:15] Mariana__: great. [13:52:03] hi, i need some of the functionality of SemanticQueryRDFS++, but it seems to be broken, at least, my smw is down, when i try to include SemanticQueryRDFS++. Is there another extension to do the transitive querying in smw? [14:29:10] hi, i need some of the functionality of SemanticQueryRDFS++, but it seems to be broken, at least, my smw is down, when i try to include SemanticQueryRDFS++. Is there another extension to do the transitive querying in smw? [14:48:43] quietschie: we all saw it the first time ;) [14:48:50] quietschie: what do you mean by "down"? [14:49:30] the same thing as if i forget to put a semicolon at the line ending in LocalSettings.php [14:50:22] what does the log say? [14:50:47] so i guess there is a php error somewhere, or it is a problem with the rights of those executable files...but i don't like the idea, that a php-file runs an executable [14:50:57] which log? [14:52:39] i would prefer having another extension instead of debugging this one, i guess, SemanticQueryRDFS++ is far beyond my needs anyway [14:53:28] is there another extension that is able to reason? [15:01:12] php doesn't need executable permissions [15:01:20] read is enough [15:01:33] you should really check the logfile if you want to debug this [15:01:49] /var/log/apache/... or /var/log/httpd/.. [15:05:22] sven: http://www.mediawiki.org/wiki/Extension:SemanticQueryRDFS%2B%2B#Download_and_Installation says, that there are executables, and php (or who else?) executes them [15:06:06] quietschie: to backtrack a little, what is the specific querying that you want to do? [15:08:31] yaron: i'm not so sure about that right now...the goal is to automatically create templates according to properties that are in the ontology of my topic [15:08:57] Could you give any more information than that? [15:10:10] all the articles are linked with a [[belongs to::superclass]] property, and i want to parse the tree to gather all the properties that any of the nodes on the way to the root-class have [15:10:53] So you want to modify pages, to add templates to them? [15:11:03] and then, i'd like to add a field of each of the properties to the page [15:11:06] yes [15:11:21] Oh - so they already have template calls, but you want to add one or more fields to each one? [15:12:04] or instead i thought of a new button "create page"that handles the template generation for a new field [15:12:21] sry, didn't understand your last post [15:12:25] But "create page" implies that the page doesn't exist yet. [15:12:37] yes, thats right [15:13:06] So you want to automatically create pages, based on semantic properties that link to those pages? [15:13:14] but a page belongs to one of the classes, e.g. is an instance of some article in my ontology [15:13:49] Oh. So basically you're talking about importing data into the wiki? [15:13:53] based on the properties that all of the superclasses have [15:14:52] What is the source of this data - is it data that's already elsewhere in the wiki, or data that's outside of the wiki, or both? [15:15:06] no, not only. The ontology knowledge is imported and used as classes, but the instances should be created using the power of the smw [15:15:28] What power? SMW doesn't create pages. [15:15:29] it is data in the wiki, or in the future wiki [15:15:58] no, but the UI, user management, semantic search etc [15:16:33] Fine, okay. (SMW doesn't do user management either, but that's another story.) [15:16:40] there are little tools that are good in modelling semantic data for domain experts [15:16:55] but mw does [15:17:35] So you want to do an ontology import... what does that have to do with SemanticQueryRDFS++? [15:17:50] nothing, i'm beyond importing [15:18:48] i wanted to use SemanticQueryRDFS++ to gather the properties of all the articles alongside to the root element [15:19:07] I'm as confused as ever. [15:19:21] sry, wasn't intended [15:19:29] Are you talking about modifying pages, or not? SemanticQueryRDFS++ doesn't modify, or create, pages. [15:19:59] perhaps i can't explain it clearly, because english is not my mother tongue [15:20:29] Just answer this question: do you want pages to be automatically modified and/or created? [15:20:47] yes and no [15:20:51] :) [15:20:56] Wonderful. [15:21:31] you don't speak german, by chance? [15:21:40] Nein. [15:21:49] :) [15:22:18] It seems like you want a few different things, and it might be better to discuss each of them individually - is that possible? [15:22:44] i'll try that [15:23:26] situation: i have domain knowledge about a domain: costumes in films that is modelled as (kind of) ontology [15:24:40] i have already imported these ontologies, preserving the tree like structure with a property: belongs-to [15:25:24] now, a new instance of lets say a hat should be created [15:26:12] therefore i want to provide a form containing all the ontology-information that is relevant [15:27:24] now, i don't want to model each property for every class, but i want to inherit properties from superclasses [15:27:57] Ah. [15:28:17] the properties can be edited step by step, so i need the forms for new instances to be created dynamically [15:29:16] and for that, i need something, that runs along the belongs-to path and fetches all the attributes, of each article on the way [15:29:29] Well, maybe - there might be other ways to do it. [15:29:37] sorry, not attributes but properties [15:29:43] About how many elements are there of this "tree"? In other words, how many different types of clothing are there? [15:29:56] german: Attribute (false friends) [15:32:37] right now 907 pieces [15:33:14] 900 different types of clothing? [15:33:50] I don't mean items of clothing - I mean types, like "hat" or "dress". [15:33:59] yes, i'm not the domain expert, but you should see how many types of shoes she knows:) [15:34:08] Man, okay. :) [15:34:32] Well, the standard way to do all of this stuff in SMW is to have a category for each type. [15:34:42] the tree has different depth [15:35:02] Don't use a property lik "belongs to", but instead have each category belong to its parent category. [15:35:12] okay, but smw was created out of the need, that categories aren't enough [15:35:29] They're not enough, but they're useful for certain things. [15:35:53] Specifically, they're used to define a class of objects. [15:36:18] ok, does that by any means infect the search behaviour? [15:36:28] affect [15:36:44] No - you can have the search cover the "Category" namespace as well. [15:37:50] So, to clarify, a page like "Red shoes number 123" would go into "Category:Shoes". [15:38:02] And then you can have a template and form for each such category. [15:38:25] ok, but what would you suggest: how can i make this traverse through the tree to the root and gather all properties? [15:38:35] If there are 900 types/categories, then it might not make sense to have a separate form/template for each one of them, though. [15:39:09] Well... ideally, the Page Schemas extension could handle that, through inheritance, but that's only a planned feature - it doesn't exist yet. [15:40:03] At the moment, I think the only real option is to have each template include all the properties that apply - no automatic inheritance would happen. [15:40:03] well i guess, it makes sense, it's the only way to make sure that there is no "unallowed" data [15:42:37] New patchset: Jeroen De Dauw; "Improved SMWDescription docs and style" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47593 [15:42:54] that hurts a bit:) [15:43:43] but i guess there has to be a way to do this, if i have to code it, i'll code it, but it would be great, if it already worked [15:44:18] Change merged: Jeroen De Dauw; [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47366 [15:44:29] Well, as I said, I think Page Schemas is the answer, going forward. [15:45:02] Any other solution would be a mess, as far as I can think. [15:45:02] where can i find something about it? [15:45:17] !e Page_Schemas [15:45:18] https://www.mediawiki.org/wiki/Extension:Page_Schemas [15:45:29] nice :) [15:48:38] New review: Jeroen De Dauw; "I am getting very confused about the dependencies now..." [mediawiki/extensions/SemanticMediaWiki] (master); V: 0 C: 2; - https://gerrit.wikimedia.org/r/47090 [15:49:37] Change merged: Jeroen De Dauw; [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/46933 [15:53:38] New patchset: Jeroen De Dauw; "Merge "smw.dataItem.factory() to enable dataItem type hinting during parseJSON()"" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47595 [15:54:30] Change merged: Jeroen De Dauw; [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47595 [15:55:24] New review: Jeroen De Dauw; "I cherry-picked this into master, fixed conflicts and merged https://gerrit.wikimedia.org/r/#/c/47595/" [mediawiki/extensions/SemanticMediaWiki] (master); V: 0 C: 0; - https://gerrit.wikimedia.org/r/47090 [15:56:25] New patchset: Jeroen De Dauw; "SMW\FeedResultPrinter to move into the SMW namespace" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47570 [16:01:50] New patchset: Jeroen De Dauw; "Fix type hints" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47596 [16:01:51] Change merged: Jeroen De Dauw; [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47570 [16:02:00] Change merged: Jeroen De Dauw; [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47596 [16:02:26] yaron: the things done in page_schemas look good, but i don't think that they match my needs. I want to avoid, that simple users or contributors have to have domain knowledge, and some things can't have some properties, even though they are in the same category...i'll think it through :) Thanks for listening and thinking! [16:03:00] New review: Jeroen De Dauw; "Needs rebase" [mediawiki/extensions/SemanticResultFormats] (master); V: 0 C: 0; - https://gerrit.wikimedia.org/r/46650 [16:03:15] This approach wouldn't require users to have to domain knowledge. [16:05:00] New review: Mwjames; "I'll this later, since a bunch of updates are coming due to the .factory() stuff" [mediawiki/extensions/SemanticResultFormats] (master); V: 2 C: -1; - https://gerrit.wikimedia.org/r/46650 [16:06:31] but it would mean to model everything...i'll think about it...Thanks a lot! [16:38:47] New patchset: Mwjames; "smw.dataItem.factory() to enable dataItem type hinting during parseJSON()" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47598 [16:39:13] Change abandoned: Mwjames; "This doesn't work ..." [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47598 [16:57:01] New review: Jeroen De Dauw; "Abandon?" [mediawiki/extensions/SemanticMediaWiki] (master); V: 0 C: 0; - https://gerrit.wikimedia.org/r/47090 [17:02:17] Change abandoned: Mwjames; "This will take me hours for a fix, so I abandon this one but really Git really -> f... m. a.." [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47090 [17:14:16] lol at last commit message [17:14:20] yaron: ^ [17:15:37] I don't know what he meant to say there, but it can't be good. :) [17:17:10] yaron: I did not make that comment :p [17:17:43] I know. [17:22:06] yaron: oh, I'm blind myself :) [17:44:11] New patchset: Mwjames; "Fix property test and add smw.dataItem.wikiPage.isKnown()" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47603 [18:05:22] Change merged: Jeroen De Dauw; [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47603 [18:12:15] New patchset: Jeroen De Dauw; "Update authors list to relect current state of the project and link to the more comprehensive list on smw wiki" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/47605 [18:14:23] New review: Jeroen De Dauw; "Congrats James! Do feel free to link your name to something. I can recommend https://www.youtube.com..." [mediawiki/extensions/SemanticMediaWiki] (master); V: 0 C: 0; - https://gerrit.wikimedia.org/r/47605 [21:34:37] Hello, all... got a heck of a question today :) [21:35:55] The rating property relationship works fine, no problems with that. [21:35:57] I recently converted a bunch of pages to have all their properties saved as internal objects so I can also have a rating on each property. [21:36:36] Now I am trying to filter pages based on properties. This used to be a few queries for properties. Now it's queries on properties of internal objects. [21:37:15] is it possible to get an overview (plans, implementations) for client side smw javascript ? [21:37:17] The problem is that one internal object has property A, and another internal object has property B. [21:38:30] I can filter on A or B... but if I filter for A and B there are never results because A and B are never on the same internal object together, so there is never a page with both properties. [21:39:06] It would seem I have a fundamental design flaw in my setup. [21:40:04] Any thoughts on how the heck I can store the data I need (properties that associate together) and still filter them like they were all one one page? [21:42:12] Frank: if the two types of internal objects are completely different from one another, why do you want to display them together? [21:42:25] ...or do you just want to display the parent pages together? [21:43:28] I want the parent page(s). [21:44:15] Hm... there may be a better way to do this whole thing, but one option is to use Semantic Compound Queries. [21:45:02] OK, I'm familiar with compound queries. I have the extension installed on the wiki. [21:45:20] Alright, cool. That should work, I think... [21:46:05] Any thoughts on a better way to approach this overall? [21:47:00] Well, for that I'd really have to know more about the data structure. [21:47:01] My gut feeling says a better design would make it easier to filter and query. [21:48:19] The data scructure isn't all *that* complex, I dont think. [21:49:10] I have 3 categories, and the pages in each category has a particular set of properties. Nothing unusual there, I imagine. [21:51:02] I use forms and template to assign all the properties. Originally I used plain semantic properties. I recently had to go back and add a rating to every property. Because some of the properties are multiple, I went with SIO so that I could always have a set of property and rating. [21:52:03] Oh... so some of the ratings are regular properties, and some are stored with internal objects? [21:52:44] At this point, 99% of my page properties have a rating. Any property with a rating is store in an internal object. [21:53:02] Alright. [21:53:33] What's an example of a property? [21:54:50] In my people category, I have a property for a web address. [21:55:21] I collect and store the web address and a rating together in an internal object. [21:55:49] Web address, meaning a URL? What does it mean for a URL to have a rating? [21:56:22] Yes, like a URL. [21:58:05] The rating would be like a letter grade, like A, B C D, F. [21:58:19] What does that mean, for a URL to get a low rating? [21:58:19] It's kind of generic. [21:59:16] The rating is more of a note about the property. [21:59:50] We wouldn't want to query by the rating, the property is the important part. [22:00:15] Well, what would it mean if someone entered a URL, and an "F"? [22:01:11] It would mean it wasn't a good site. An A would indicate a very good site. [22:01:37] Oh - but isn't the entire page, in this case, about a specific site? [22:01:54] Kind of like a rating with stars. That's a much better analogy. [22:02:33] Exactly. The rating describes the URL, not the subject/wikipage [22:03:06] I don't understand - let's say the page is called "Google", and one the properties is "Web address", with value "http://google.com". [22:03:19] Why would the rating apply to the URL, and not to the whole page? [22:03:36] I just don't understand what the use is for so many ratings. [22:05:23] I know, it's a lot of ratings. It's not my choice to have a rating on everyhting. It's for the government :) [22:06:09] So is it fair to say that most of these ratings are meaningless? I don't mean to sound rude, I just want to fully understand the situation. [22:07:34] No, not rude at all. It's mostly meaningless, I just have to keep them together if it's there. [22:07:51] Alright. [22:08:10] So - what is it that you want to display with your query? [22:11:18] I had queries that acted like filters, so you could filter for just people with a website, or just people with certain interests (cooking, gaming, sports, etc), or a person with a website and certain interests. [22:12:22] Because all the properties are in seperate internal objects, it's going to be tricky to look for more than one thing at a time. [22:12:32] Okay. Well, that doesn't seem so hard, actually. [22:13:03] If it's "A and B", just combine the parts in a query. If it's "A or B", use Semantic Compound Queries. [22:14:26] I tried a pretty simple "A & B" type query, and I got results for one property or the other. [22:15:02] For instance, PersonA is into cooking and has a website. [22:15:19] Oh, yeah... "A and B" is a little complex in this case. [22:15:50] Yeah, I guess there's actually no way to do it. :( [22:16:19] I query for interests in cooking, and I get PersonA#1. I query for somebody with a site, and I get PersonA#2. [22:16:26] Oh, wait never mind. [22:16:48] New review: Kghbln; "Congrats to James! See my inline comment." [mediawiki/extensions/SemanticMediaWiki] (master); V: 0 C: -1; - https://gerrit.wikimedia.org/r/47605 [22:17:00] I query for both, and I get nothing, because There's no one page with both properties... but I keep thinking I just need a new approach. [22:17:42] Here's how you do it: {{#ask:[[-Is part of page.Has interest::Tennis]][[-Is part of page.Has interest::Golf]]}} [22:18:46] Oh, that looks awesome. I'm not familiar with "-Is part of page" [22:18:58] Well, I made up the property names. [22:19:15] But the "-" is important - that makes it an inverse query. [22:21:55] So "Is part of page" is the name of the interal object? Or is that a keyword in this case? [22:22:28] Internal objects don't have names - it's whatever the main/first property of the #set_internal call is. [22:22:43] All your #set_internal calls should have the same first property, in this case. [22:23:00] You're right, I think of that as a name but it's not. [22:23:45] Alright - yes, it's a little confusing because for #subobject the first parameter is the name. [22:24:07] So if I am using #set_internal:Interests [22:24:24] I would [[-Interests.Has interest::Golf]] [22:24:30] Yeah. [22:25:15] You might want to use a different property name, but you don't have to. [22:25:23] Thank you, I appreciate that you've been so patient with me today. I'll give this a shot and let you know how it goes. [22:25:39] Cool. [22:33:59] ok my turn. is it possible to get an overview (plans, implementations) for client side smw javascript ?