[19:27:55] tech talk in 30 min [19:27:58] info in topic [19:56:39] Tech talk starting in 4 min [19:58:35] we're live [19:59:32] Yay tech talk! [19:59:46] the hangout is live but we have not started [20:00:13] starting in a moment [20:00:32] you can ping me with questions if you have them for the speaker and I can ask [20:01:17] starting! [20:01:19] http://www.sebastiankarcher.com/linklist will be useful if you're following! [20:01:34] we have 7 people on the hangout, 5 people in SF and 11 watching on youtube [20:04:11] questions at any time OK - [20:07:03] Mvolz is fabulous. [20:07:34] Word! [20:09:12] those looks like beamer-based slides. [20:09:29] they are, subbu! Marko can tell you more about them later. [20:09:40] also, I'll upload them when I can on Commons. [20:10:37] Elitre, i am familiar with beamer .. my april 2014 tech talk also used beamer for the slides. [20:11:20] sure - I think Sebastian had a particular tips about those which surprised Marko earlier :) [20:11:41] subbu: actually, sebastian explained those are beaner slides generated out of md [20:11:43] ah, ok. i am curious about it now. [20:11:51] out of md? [20:11:57] s/beaner/beamer [20:12:01] subbu: markdown [20:12:24] ah, markdown. ok. yes, i have seen that. [20:13:09] slides now at https://commons.wikimedia.org/wiki/File:Translators-citoid.pdf . [20:13:22] cool thnx Elitre! [20:18:03] I'm discovering that I can fix Le Monde's Zotero xpath and solve a problem on fr.wp. So cool! ^^ [20:18:08] Are we going to have a demo? [20:18:28] Trizek: I don't know, but knowing you can do that now is sooo cool. [20:18:35] gr8! [20:19:18] le monde's my fav newspaper (especially M le magazine) [20:20:19] Maybe also interesting for you is the still open issue about French newspapers: https://github.com/zotero/translators/issues/826 [20:20:27] mobrovac, i assume you've asked the xulrunner qn.? [20:21:09] subbu: "the question" :P not yet [20:21:30] PhilippZ: what about that in particular? [20:22:13] I was wondering if there are any plans to standardize the licensing of translators (e.g. by chasing down contributors and ask for permission to relicense). [20:22:13] ATM, you can't use Le Monde on fr.wp the way we are used to, mobrovac, because of a bad path on Zotero. [20:22:16] does he know more about breaking the xulrunnder dpeendency? [20:22:22] Elitre, rfarrand ^ [20:22:26] It's now GPL/AGPL or unlicensed? [20:22:42] i.e. zotero depends on xul runner which is apparently not maintained anymore. [20:22:46] how do you pronounce xulrunnder? [20:22:53] xul runner sorry. typo [20:22:56] subbu ^ [20:23:08] zuul .. i think. [20:23:10] Rintze: I'll ask your question after Subbu's one. [20:23:25] OK, cool, you are asking Elitre ? [20:23:26] :) [20:23:42] Sure. [20:24:07] I can't see it very well. Zooming in might help. [20:25:04] One quick try is appreciated, but if it doesn't work, then that's okay. [20:25:04] Maybe lower screen resolution? [20:25:07] sorry whatami, we tried. [20:26:04] That's very easy to read. [20:26:48] Sebastian you can try to open chrome://scaffold/content/scaffold.xul in firefox and make it large with Ctrl + [20:27:05] I'm not that sure I'll fix Le Monde problem. Code. So much code. :/ [20:27:13] Elitre: i can ask the xulrunner q since i know what subbu wants :P [20:27:27] great, thanks mobrovac! [20:27:51] i want what mobrovac wants. [20:27:57] lol [20:28:12] subbu: you want a glass of wine? :D [20:28:26] sure. :) [20:31:48] Trizek: What is the problem with Le Monde at the moment? Seems to work for me in Zotero... [20:34:24] PhilippZ: The problem is linked to the visual editor. Go to a sandbox on fr.wp (brouillon, on top of the page), open in with VE, click on Cite > Automatic, paste that link http://www.lemonde.fr/sciences/article/2015/05/22/le-chien-meilleur-ami-de-l-homme-depuis-plus-de-27-000-ans_4638417_1650684.html and click on generate. Date is mising. [20:34:27] *missing [20:34:46] Do that on my sandbox if you need https://fr.wikipedia.org/wiki/Utilisateur:Trizek_(WMF)/Brouillon [20:38:31] we have that in dom libraries like domino [20:39:22] cookies .. not in domino. [20:40:14] Do you know if https://github.com/proquest/PME is still active? Would it circumvent the XUL-dependency? I'm not sure how it works. [20:40:19] subbu: we already have cookies handling in citoid [20:40:24] Trizek: This website works for me in Zotero and gives: "date": "2015-05-22T09:40:40+02:00" [20:40:30] but Mvolz mobrovac do we need cookie emulation? [20:40:30] all that is needed is anonymous read. [20:40:57] There are individual websites that won't accept the request without the cookie. [20:40:57] (also, Mozilla seems rather keen to get rid of XUL in the next year or two, in which case Zotero would have to be recast onto another platform like node.js) [20:40:59] Cool, PhilippZ. that may be a problem on Template definition on Fr.wp. [20:41:14] Thanks! [20:41:28] Subbu, if memory serves, Jstor requires a cookie. [20:41:48] oh, hmm .. interesting. [20:41:58] unfortunately most scientific sources require cookies subbu [20:42:04] mobrovac, is there a follow-up question for later? [20:42:45] Rintze: hope that helped :) [20:43:01] Elitre: possibly, i was wondering if there were similar attempts to "port" zotero to nodejs [20:43:03] Yes, thanks. [20:43:09] i [20:43:13] i'm not aware of any [20:43:49] may we get the video in full view? [20:43:57] The Zotero codebase is rather huge, so I don't expect anybody to port it other than the core team itself. [20:44:33] There was an attempt to write translation-server in node originally but it was abandoned. It's on github. But still uses Zotero standalone anyway. [20:44:45] Rintze: my current understanding is that the crucial part is to make ZU work in node [20:44:58] jeblad: we tried and failed, he's showing relevant code now and then in his notepad though. [20:45:14] ZU? [20:45:21] Yes or you'd have to write all the ZU functions from scratch. [20:45:35] Zotero Utilities library? [20:45:49] ya [20:45:55] Ah, okay. [20:46:02] mobrovac: There is a server side api for a lot of functions of zotero, if you need more than the translators. [20:46:14] and also this Zotero.newItem fcn although that looks pretty simple? [20:47:00] PhilippZ: what we really want is to get rid of xulrunner :) [20:47:45] Mvolz: without looking at the code, i'd say it's just a stub giving back an object prefilled with empty fields for the given article type [20:47:57] ya [20:48:31] .complete is a semaphor though? [20:48:41] Mvolz, mobrovac how many translators does zotero have? [20:48:57] I think Sebatian said 300 or so [20:48:58] subbu: 300+ [20:49:15] Mvolz: looks like .done() in mocha [20:49:19] too many to port out of zotero in a standalone framework. [20:49:19] https://www.zotero.org/support/translators says 300 [20:49:31] Do you know if Zotero have an important ISBN database? [20:49:39] but I do want to ask why the complete list has been broken for months now :/ [20:50:00] do you mean the tests Elitre? [20:50:40] It looks like a basic translator is approximately half an hour's work. [20:50:42] "You can find a full, automatically generated list of all Zotero translators here. The list is generated daily and includes automatic tests of most translators, so that Zotero developers and volunteers can quickly identify and repair issues." [20:50:51] hmm .. http://zotero-translator-tests.s3-website-us-east-1.amazonaws.com/ .. no tests run recently. [20:50:57] but that page doesn't include any such list :/ [20:51:40] Any other questions? We have nine minutes left. [20:51:43] last chance for questions, we have 9 min [20:51:48] hah :) what whatami said :D [20:52:38] what is the process to include a specific translator in citid? [20:53:09] jeblad, that's probably a question for mvolz or mobrovac ? [20:54:08] first submit it to sebation on github :) [20:54:33] Is there anything exciting on the Citoid roadmap? (other than getting rid of XULRunner?) [20:54:49] wil ask that, Rintze! [20:56:09] actually, that's also for mvolz and mobrovac :D [20:56:24] Trizek: Zotero has no ISBN database itself, but it can ask LoC, Worldcat, GBV as far as I know. [20:57:13] I'll copy the linklist to a mediawiki page tomorrow. [20:57:47] For REASONS we cannot use Zotero's access to worldcat :P [20:57:54] PhilippZ: Thanks. It is something that people would appreciate: generate citations from an ISBN. Some users are populating Wikidata for that purpose. [20:58:21] Trizek: yeah, has been a big issue, I think wikidata could solve this [20:58:31] eventually :) [21:00:02] jeblad: https://github.com/zotero/translators is where you should submit new translators [21:00:29] these periodically get merged into citoid's repo [21:00:38] Nothing additional to that? [21:00:41] and they go through review from the Zotero team [21:00:48] Nobody to nag? [21:00:52] you just have to make sure they have the 'v' flag [21:00:53] jeblad: Shout at me if it's not updated in a while. :-) [21:01:01] :D [21:01:47] okey,.. must take a new look [21:02:30] jeblad: the point is that currently there are no wp-specific translators [21:02:43] Did start on a translator for nb.no, but there is a xml-based interface and didn't figure out how to switch the url [21:03:23] the site nb.no is a huge kind og Google Books with complete texts for Norwegians only [21:03:58] The metadata is open and can be reused [21:04:58] no.nb should work OK, yes -- might be worthwhile seeing if they have an API rather than scraping of the page [21:05:02] jeblad: as skarcher said in the talk, a good approach would be to do a first PR against the translators github repo and work with more experienced zotero coders to get to the final result [21:05:43] http://www.nb.no/services/search/v2/ [21:06:04] jeblad: xpath works with xml too [21:06:15] Not sure, but I think we can do a ISBN lookup through the site and reuse the metadata ... at least we was told we could reuse BIBSYS-data for Wikipedia a few years back [21:06:49] mobrovac: yes, I know [21:06:56] yes, so I think that'd be the way to go: make an API request, then get the XML back and read that into Zotero [21:07:25] check out e.g. the arXiv translator, which does that (it also looks up by DOI, so don't get confused by that) [21:07:47] Usually we will have the URL to the web page, then we must rewrite that and get the xml page [21:08:20] yes, that should work -- or you can find the id for the API request somewhere on the page [21:08:29] ok, I'll take a look on that... [21:08:33] any other sites people are working on? nb.no certainly looks nice [21:10:33] http://bibsys-almaprimo.hosted.exlibrisgroup.com/BIBSYS:default_scope:BIBSYS_ILS71471972410002201 [21:11:10] This is a book from BIBSYS [21:12:00] that should work. ExLibris Primo catalogs generally do [21:12:29] OK, I have to run -- you can always find me on the zotero dev group, so any more questions point there. Thanks! [21:12:58] Yes, we get some of the data, but often it is messed up [21:13:53] I have an additional uestion about citoid [21:14:25] jeblad: Mvolz is the expert but I can answer too. :-) [21:15:15] I've asked retriever-info.no / atekst.no about access to metadata to get citoid working for news articles [21:15:29] Thei sell access to old news articles [21:15:38] They, .. :( [21:16:25] But to let someone get access they need a fixed address for access [21:16:49] I guess citoid sits on a fixed address at WMF, or..? [21:17:49] jeblad: Fixed IP? I believe they'll come from an effectively-fixed IP inside the WMF cluster, but it's not guaranteed to be fixed. [21:18:08] IP-range? [21:19:29] I guess the question is whether WMF controls the range or if it sits on an server farm. [21:19:53] If its not an fixed address someone could gett free access, which is bad. ;) [21:20:04] It'll come from the WMF server cluster. [21:21:09] ok, I'll ask them again whats possible [21:21:30] jeblad: I.e., from 198.73.209.0–198.73.209.255 but that would also cover WMF Labs. [21:21:33] It would b really nice to have citoid for old news articles [21:21:46] would be.. [21:21:49] jeblad: Thanks! Great if you could raise a Phabricator ticket to track this. [21:22:22] I'll send them an email [21:24:43] Thanks