Wikivoyage:Script nominations

From Wikivoyage
Jump to: navigation, search

According to the Wikivoyage script policy, scripts have to be approved by the Wikivoyage administrators. To create a script that runs against Wikivoyage, post the name and reason for the script beneath the line below.

Explain why we need the script, why it can't be done by hand, and what the script will do. If 2 administrators voice their support for the script and there are no unresolved objections, the script can be run with a bot flag. If objections arise later, the bot flag can be undone.

Scripts that have passed through this process can be found in Project:script nominations/Archive.

NOTE: you must apply for approval on each language version of Wikivoyage. Approval on this page only allows you to run a bot on Wikivoyage in English.

PoI File[edit]

This is more a testing-the-water at the moment and comes from a question I asked in the Travellers' Pub. Given many articles are starting to get latitude and longitude tags in listings, and given the popularity of GPS (not just SatNav devices but many mobile phones are starting to have GPS built in) would there be any use in creating a script that can scour a page for listings containing lat/long tags and creating a PoI file for that page?

I would propose to use the GPX Standard [1] since it seems to be the best attempt at a common format (and I know Garmin MapSource will open these files). Just then a case of how to present it to the reader - do we have a /PoI page for each article that shows the raw XML of the GPX file that the reader then has to copy and paste into a text file (not my prefered way) or is there perhaps some way to generate a file that can be downloaded straight away as a GPX file.

Anyway, if people think it would be of use and would fit in with policy (I suspect it should since it's generally speaking a read-only script where the article pages are concerned) and also can point me to some useful articles or examples on how to create scripts for MediaWiki I would be interested in giving this a go. (WT-en) Nrms 03:07, 8 March 2010 (EST)

If we use a template for coordinates that emits a geo microformat, as proposed at Template talk:Listing#Coordinates (and as done on Wikipedia by Template:Coord) we can then use a one-per-page template like en.Wikipedia's Template:GeoGroup to make them available as a downloadable KML file, as used, for instance, by Google maps (and by conversion, as a GPX file) in real-time. I could add the necessary code in about 10 minutes, and as only HTML classes are involved, there'd be no processor overhead. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:57, 25 November 2012 (UTC)
Only a fraction of our listings are currently templatised and very few currently have co-ordinates. That will need to change if we are ever to generate locator maps or links to auto-generate same. Using the toolserver script would require that script change to accept voy: links instead of only wikipedia: pages. If a listing had the tags or templates (so that address, name and other portions can be automatically extracted) it might be possible for a 'bot to obtain (lat, long) from an external source and insert it into the article. In some cases, though, we don't even have that... the listings are plaintext. A locator map would be invaluable but there's a lot to be done before we can do these for all pages. K7L (talk) 16:26, 29 November 2012 (UTC)

Snowbot[edit]

I'd like to run Snowbot on here to replace old reference to shared with local pages where applicable, convert links from external to interwikis, etc, automatically to help with the migration process and cut down on those red links. I have run bots on enwiki and other wikis (including adminbots) many times without many issues. Snowolf How can I help? 01:37, 12 November 2012 (UTC)

  • Support. Provided there are no objections, given the large amount of cleanup needed I'd suggest this be approved quickly. -- Ryan • (talk) • 02:20, 12 November 2012 (UTC)
  • Speedy supportcacahuate talk 04:33, 12 November 2012 (UTC)
    • 04:47, 12 November 2012 Wrh2 (Talk | contribs | block) changed group membership for User:Snowbot from (none) to bot (set bot flag)

Thehelpfulbot[edit]

From the user page, User:Thehelpfulbot seems to be a bot, and since it is updating country flags I've tagged it as such. Listing it here per policy. -- Ryan • (talk) • 19:22, 12 November 2012 (UTC)

  • Support. Looks like another find-and-replace bot, which is particularly useful right now. -- Ryan • (talk) • 19:22, 12 November 2012 (UTC)
  • Support Per above, no problems Addshore (talk) 23:09, 10 February 2013 (UTC)

The Anomebot2[edit]

Swept in from the the pub: K7L (talk) 21:33, 25 November 2012 (UTC)

Gocoding bot Hi! I'm a member of the geographic coordinates project on the English-language Wikipedia. I operate w:en:User:The Anomebot2, a bot that adds and maintains geocodes on that project. You can see some of its recent work here.

I've extracted a list of articles on en: Wikivoyage that currently do not have a Geo template, but for which coordinates are available on the corresponding article on enwiki. I've done quite a lot of dataset cleaning, and I have a list of 2317 6687 articles currently without geodata templates for which I have coordinates from enwiki. If someone would like to give User:The Anomebot2 bot rights here, I'd be happy to add them to these articles. -- The Anome (talk) 21:21, 25 November 2012 (UTC)

  • Odd that no one's commented on this... current status quo is that co-ordinates are displayed for cities/towns if we have them (although the lat= and long= fields in individual listings currently do nothing). As such, this looks like it would be useful. K7L (talk) 20:16, 27 November 2012 (UTC)
  • Support. This seems straightforwardly useful. I'll give this a day and then flip the switch. --Peter Talk 20:20, 27 November 2012 (UTC)

Flagged --Peter Talk 00:35, 29 November 2012 (UTC)

Many thanks. I'll create a tweaked version of the bot back-end designed to handle Wikivoyage's article format, then start to make edits later this week. -- The Anome (talk) 19:13, 2 December 2012 (UTC)
Update: I've now added coordinates to 10995 11000+ articles, which is as many as I could do automatically for the time being without getting into unreasonable amounts of programming work writing more pattern-matching code. -- The Anome (talk) 22:40, 7 December 2012 (UTC) -- updated 2012-12-09

Additional generation of interwiki links to Wikipedia[edit]

Following my recent addition of 11000+ coordinates sourced from equivalent articles on enwiki and/or GNIS, I noticed that there are still lots of articles here that could be interwiki linked to Wikipedia, but aren't. I can auto-generate a lot of these, with high accuracy, on the following basis:

For locations outside the U.S.:

  • contains isPartOf or isIn template on Wikivoyage
  • same basename on both Wikivoyage and enwiki
  • one and only one article on enwikivoyage with that basename (i.e. after stripping of disambiguation suffixes, and, in the case of Wikivoyage, also slash-prefixes)
  • same for enwiki, for the same basename
  • the single country location identified by tree traversal of the Wikivoyage breadcrumbs graph agrees with country identified by tree traversal of the Wikipedia category graph
  • NGA GNS data indicates that there is one, and only one, place of that name in that country

I can also generate links for U.S. articles using by-state-and-county disambiguation (using GNIS data) for article names of the format of Wikivoyage "Placename (State)" <-> Wikipedia "Placename, State". (or just Wikipedia "Placename", using the same rules as for non-U.S. locations, but using GNIS data instead of GNS.)

Once that's done, I can trivially add the resulting interwiki links to the articles here. My best guess is that this will generate at least several thousand new interwiki links. These will, in turn, act as a data source later creation of backlinks from enwiki to Wikivoyage, copying of coordinates (where appropriate) from enwiki to Wikivoyage, and finally eventually integration of all of these with Wikidata, OpenStreetMap etc.

If my bot can be authorized to do make these changes, I can put this on my queue of things to do in the next week or so. -- The Anome (talk) 14:00, 9 December 2012 (UTC)

Sounds good to me. Pashley (talk) 14:48, 9 December 2012 (UTC)
My support too. --Globe-trotter (talk) 14:53, 9 December 2012 (UTC)
  • Support. The ability and willingness of Wikipedia users like yourself to help out with scripting these sort of tasks is hugely appreciated and extremely useful - I find it extraordinarily impressive how much work is done behind the scenes when I view Special:RecentChanges with the "show bots" flag enabled. -- Ryan • (talk) • 16:47, 9 December 2012 (UTC)
  • Comment: If you do ever use this to generate backlinks to Wikivoyage from the 'external links' on Wikipedia, could you please ensure that anything labelled "stub" or "outline" here does not get a backlink? Most of these are "X is in Y", followed by an empty template, and are utterly worthless. (No idea why we keep them.) K7L (talk) 15:57, 11 December 2012 (UTC)
Thanks. I'm on the job -- as before, it may take several days to a week for me to do this. -- The Anome (talk) 18:44, 9 December 2012 (UTC)
Update: I've now generated 4000+ potential interwiki links for non-U,S. locations. About half of these have interwiki tags already, so expect a gain of about 2000 interwiki tags after this pass has finished. Then I will do U.S. locations. -- The Anome (talk) 15:03, 10 December 2012 (UTC)
OK, that's it for now. the total to date is 11170 coordinates added, and 2150 interwiki tags. If you can think of any other ways I could help, please drop me a line at my enwiki talk page. -- The Anome (talk) 19:26, 10 December 2012 (UTC)

OK, I've got another couple[edit]

I've been poking around to find small annoyances, and seeing if any of them can be easily fixed using a bot. Here's another one.

547 pages currently have IsPartOf tags that point to pages that have been renamed. An example is Córdoba_(city,_Argentina) which is currently marked as IsPartOf Cordoba (province, Argentina), but should be marked up as IsPartOf Córdoba (province, Argentina) instead. Similarly with [Polesie]], marked as IsPartOf Lublin_Voivodship instead of the correct Lubelskie.

As a result, the breadcrumbs in these articles don't work correctly. I can fix the lot using a script to update all the broken tags to point to the correct redirect targets, if you approve.

Another thing I can do is track down the last remaining IsIn tags, and fix them, too. -- The Anome (talk) 02:18, 11 December 2012 (UTC)

Fixing broken breadcrumbs has been on a lot of people's TODO lists and has never gotten done, so if you can fix it that would be awesome! -- Ryan • (talk) • 02:25, 11 December 2012 (UTC)
How does this work? Does it find "isIn" or "isPartOf" tags pointing to a redirect with no breadcrumbs? I'm just wondering if the kludge Russia (Asia) (a redirect which contains #REDIRECT [[Russia]] {{isPartOf|Asia}} will be left alone (the redirect is a trick to allow Siberia "isIn" Russia (Asia) so that it gets Asia breadcrumbs instead of Europe. K7L (talk) 03:01, 11 December 2012 (UTC)
It simply looks for IsPartOf tags that point to a redirect page, and uses that substitution, without looking to see if the redirect target itself has a tag. It won't necessarily make things right, but it should usually make them better, and never make them worse. I could add the check, but I can't see the point: simple heuristics which get things 95% right, and don't make anything worse, are generally preferable to more complex heuristics which are harder to debug. I don't currently try to follow multiple redirect chains: although I easily could -- it's a one-liner, and I can't see how it could be harmful.
I would leave the Russia (Asia) kludge alone, as it points to a page fragment, not a page: it does, however, expose an edge case for the whole breadcrumbs system: the assumption that the graph of inclusion of territories and geographic boundaries forms a tree, as opposed to a more general graph, which isn't easily fixed, except programatically as part of a rewrite of the breadcrumbs support code.
Fixing IsIn is even simpler: I will just grep them all out of the dump, and unescape their target string at the same time as changing IsIn to IsPartOf. Again, I won't do any over-elaborate checks while doing this, such as checks that the target exists, or has a tag, since it wouldn't actually make anything any better: experience has shown me that it's better to break things into a set of simple, dumb, do-no-harm tasks, than to try to resolve everything in one pass.
By the way, can I have rollbacker rights for this parent account, please? It would make bot debugging that much easier. -- The Anome (talk) 13:57, 11 December 2012 (UTC)
Could you also change all IsIn references to IsPartOf? --Globe-trotter (talk) 21:40, 11 December 2012 (UTC)
Yes: I mention it above. -- The Anome (talk) 22:30, 11 December 2012 (UTC)
I don't think we have the rollbacker right enabled on this wiki. sumone10154(talk) 18:51, 12 December 2012 (UTC)

OK, I've now replaced all 6300-or-so instances of IsIn in pages in the main and File: namespaces with the appropriate IsPartOf tags. Uses in meta pages remain untouched. Yes Done

And also fixed nearly all of the instances of IsPartOf pointing to redirect pages. There are probably a few edge cases remaining: I've been cautious in doing these. Yes Done -- The Anome (talk) 14:25, 14 December 2012 (UTC)

DMOZ[edit]

I'm working on generating a new set of DMOZ links, as the existing links here seemed to be in rather poor repair: are people here OK with me doing this? -- The Anome (talk) 19:28, 14 December 2012 (UTC)

Sumone's bot (below) previously attempted to fix some of them up by removing the incorrect "Region" prefix. What is your bot doing? And assuming it's nothing insane I'm completely fine with it - you've done great work thus far. -- Ryan • (talk) • 19:43, 14 December 2012 (UTC)
Scanning the DMOZ, Wikivoyage and Wikipedia and GNS category trees, and generating links for everything that I can find that is unambiguous. The matching logic is really quite simple (traverse all root-to-leaf arcs in all three trees, checking for each top-to-bottom paths for same name, up to disambiguation, same country, and only one of them in that country in all three sources) and I've got all the necessary datasets and tools in place already. It may well be that Sumone has found and/or fixed all of these already, in which case this will just validate their work and provide a cross-check. If not, I will at least be able to add a few more links, or generate a list of discrepancies between my links and Sumone's, or both. -- The Anome (talk) 20:02, 14 December 2012 (UTC)
Sounds good to me... unleash the bot. -- Ryan • (talk) • 21:24, 14 December 2012 (UTC)
It looks like in some cases the bot may be adding duplicate Dmoz tags: [2]. -- Ryan • (talk) • 21:55, 14 December 2012 (UTC)
Yes, I found (and, hopefully, correctly fixed) that bug about ten minutes ago: I've made a list of the 250-or-so articles processed before that point, and I'll do another pass on just those articles to remove the duplicates: my best guess is that there are about 20 or so such articles for it to track down and fix. -- The Anome (talk) 22:07, 14 December 2012 (UTC)

Update: it looks like there are about 4000 new DMOZ tags on their way: I won't know the exact number until the bot finishes making the edits, since the last phase of the process peeks at the live copy while editing it, and performs final checks. -- The Anome (talk) 01:21, 15 December 2012 (UTC)

The bot has now added DMOZ links to 5175 articles. Yes Done -- The Anome (talk) 00:36, 16 December 2012 (UTC)
Fantastic. --Globe-trotter (talk) 01:09, 16 December 2012 (UTC)
Agreed, that is awesome. Thank you! -- Ryan • (talk) • 01:58, 16 December 2012 (UTC)

Removal of nonexistent interwiki links within Wikivoyage[edit]

There are lots of bogus interwiki links between different Wikivoyage languages, such as for example a link to fi:Rio Grande do Sul from Rio Grande do Sul. There are also large blocks of commented-out and usually completely invalid interwiki links in many articles, presumably once added as part of a templating process, that seem not to serve any purpose any more other than to generate confusion. I'd be happy to work my way through the articles removing both of these. -- The Anome (talk) 21:32, 17 December 2012 (UTC)

Support from me with one question - if Finnish Wikivoyage re-launches, is there any way to rebuild interwiki links? How do other WMF projects manage this when a new language version launches? -- Ryan • (talk) • 21:37, 17 December 2012 (UTC)
I would wait, since we expect most of the language versions to relaunch. The commented-out potential interwiki links were a bad idea, though, and should be removed. --Peter Talk 22:07, 17 December 2012 (UTC)
OK. I can easily remove just the commented-out interwiki links from the ~1500 articles containing them, while leaving the other interlanguage links in, regardless of validity (except for those linking to obviously impossible targets such as the empty string or FULLPAGENAME). I will do this in the next day or so. -- The Anome (talk) 11:52, 18 December 2012 (UTC)
Yes Done -- The Anome (talk) 23:30, 5 January 2013 (UTC)

Removal of wts: links[edit]

wts: category links serve no purpose any more, and are just clutter now: they're not even useful for finding potential commons links (something which can be achieved better algorithmically) so I see no reason to keep the wts links. I can remove the lot of them very simply, given the OK here: they're currently in ~1400 articles. -- The Anome (talk) 23:33, 5 January 2013 (UTC)

  • Support. -- Ryan • (talk) • 00:01, 6 January 2013 (UTC)
  • On other Wikivoyages I used the wts: links to find the right category on Commons, and worked in most cases very well. So the statement that it is just clutter is something I do not agree with. And only 1400 pages is not that much. I would prefer that someone would add Commons links to the articles first, before the suggested way of adding Commons links algorithmically does make the wts links not needed any more. I have experience with adding Commons links and based on that I really want to see first the method of adding those links. Romaine (talk) 02:10, 6 January 2013 (UTC)
  • I'd think that one would have a better chance of finding potential commons: links by following the wikipedia: link than by trying to match wts: ? K7L (talk) 04:02, 6 January 2013 (UTC)
    • That's right. Following earlier bot runs, most articles now have links to the en: Wikipedia. Using data taken from dumps, I can now cross-reference these links with links to Commons categories from the corresponding Wikipedia articles, to create links to Commons categories to add here. -- The Anome (talk) 10:13, 6 January 2013 (UTC)
    • OK, I'm halfway through that process now, and working on cleaning and validating the data. I currently have 6300+ commons category link candidates, which I hope should more than make up for the loss of the wts: links. More soon. -- The Anome (talk) 13:55, 8 January 2013 (UTC)
      • I'm currently adding commons: links: I will leave the wts: links in for now, until consensus is reached here about what to do with them. -- The Anome (talk) 22:23, 8 January 2013 (UTC)
        • I'll reiterate my support for removing the "wts" links - they aren't a reliable way to generate commons: links, and if your bot has already generated commons: links then they most definitely just clutter now. -- Ryan • (talk) • 22:27, 8 January 2013 (UTC)
          • I've now added links to Wikimedia Commons to 5400+ Wikivoyage articles. Yes Done

            I propose that I now set the bot to remove the redundant wts: links from all the articles now containing links to Commons. I can then do an survey of those wts: links that remain in other articles, to see whether they have any useful information content. -- The Anome (talk) 19:15, 9 January 2013 (UTC)

        • Blow them back to Allah or the deity of your choice, they're useless now. K7L (talk) 19:32, 9 January 2013 (UTC)

Update: I've removed the wts: links from pages that now have commons: links. After this, only ~700 pages are left with wts: links in them-- The Anome (talk) 20:43, 10 January 2013 (UTC)

Update: I've now removed wts: links that are exactly match the page title, and thus have no extra information content that is not already present in the article title. After this, only ~170 wts: links remain. -- The Anome (talk) 21:12, 10 January 2013 (UTC)

Update: I've now replaced the remaining 170+ wts: links with uses of the {{wts}} template, which has no visible effect, apart from transcluding the article into a hidden category Category:Pages with old wts links. I thinks this probably sorts things for now as far as wts: links are concerned. Yes Done

Coming next: a more thorough matching of wv: articles to commons and Wikipedia entries. But probably not for a week or so. -- The Anome (talk) 21:54, 10 January 2013 (UTC)

  • Update: the last remaining wts: links have now been removed manually, with most of them being converted to relevant Commons links. Yes Done -- The Anome (talk) 19:46, 11 January 2013 (UTC)

Creation of categories[edit]

I'm happy to go over the entire breadcrumb tree (which I have already parsed) to generate all the needed categories. I will initially call them all "Places in <X>". The {{IsPartOf}} template can then be alterned to automatically include articles in their respective categories. This will greatly reduce the number of pages to be edited, as otherwise I'd have to edit every single guide content page.

Can I assume this is OK? Regarding policy, we seem to already have consensus, and qualified permission, elsewhere on this page (see the K7Lbot discussion below) that this is a good idea. I can have this done quite quickly, given the go-ahead. -- The Anome (talk) 21:01, 15 January 2013 (UTC)

Note the discussions below at User:K7Lbot. There is already a tested format for categories, naming and content and the syntax of the ispartof template. Tested and works, just need to generate the mass of categories. --Traveler100 (talk) 21:44, 15 January 2013 (UTC)
Support. Using an existing 'bot for this would avoid the need to create a new one. I have a list of categories (from a database dump) on user:K7L/isPartOf but these will need to be cut down to remove any leaf nodes (a destination with nothing else under it doesn't need an eponymous category). I'd suggest the output be worded not as "X is in Y" but {{some template|X|Y}} so that every page can be changed in format just by editing some template. K7L (talk) 21:49, 15 January 2013 (UTC)
I'm working my way through the breadcrumbs contained in the dump, checking for consistency and sanity before getting ready to start creating some categories. It looks like roughly 3000 will be needed. -- The Anome (talk) 11:28, 17 January 2013 (UTC)
OK, doing them now. See Category:Val di Chiana for an example. -- The Anome (talk) 21:49, 19 January 2013 (UTC)
As it is not creating the categories top down (which I guess would be very difficult to program) I suggest once complete a minor edit to {{RegionCat}} to resave the categories and get the content updated. Or is there a better method? --Traveler100 (talk) 21:56, 19 January 2013 (UTC)
I've just added empty HTML comments to both {{IsPartOf}} and {{RegionCat}}. This should cause effectively all of the articles to be re-rendered and all the categories populated, right down to the article level. The downside is that finishing this batch of re-renders will take the site many hours to complete, but we have plenty of time -- tomorrow is soon enough to see the results. -- The Anome (talk) 23:33, 19 January 2013 (UTC)
Initial look through, looks good. Nice job.--Traveler100 (talk) 06:07, 20 January 2013 (UTC)
This is already identifying many outlines for individual destinations which were placed "isPartOf" a province/state/country instead of in a more precise subregion. Most have very little actual content. [3] is still reporting a thousand jobs in the job queue, so that should keep the server busy for a while. K7L (talk) 08:18, 20 January 2013 (UTC)
Absolutely. The output is only as good as the input. Fortunately, changing the {{IsPartOf}} entries for those articles should also change the category that they belong to, and, for those with the relevant geographical knowledge, the category tree now provides a quick way of finding article with this problem. -- The Anome (talk) 06:41, 21 January 2013 (UTC)

Creation of categories now complete. Yes Done -- The Anome (talk) 12:11, 23 January 2013 (UTC)

Awesome, thank you. This looks like it will be extremely useful for keeping the article hierarchy organized. -- Ryan • (talk) • 15:56, 23 January 2013 (UTC)

Districts[edit]

Your bot has been adding many IsPartOf tags in District articles like Shanghai/Bund, One example [4]. These should not be needed. See Wikivoyage_talk:Breadcrumb_navigation#Districts and Wikivoyage_talk:Breadcrumb_navigation#Districts_2.

Is this a deliberate work-around for the parsing bug? If not, then adding tags that would not be needed if the bug were fixed and that hide the buggy behaviour in the meanwhile is not a good idea. Pashley (talk) 17:05, 27 January 2013 (UTC)

At the rate code is being updated, it looks like bugs for which patches were written in December (like putting rel="nofollow" on WT links or redirecting tag output via the {{listing}} template) are still awaiting deployment and should go live in the next update (likely Feb 11 or so) of the actual site. The isPartOf/district bug had been reported but no fix has yet been proposed, so there's likely no realistic prospect of getting a patch through code review in time for the next update. Certainly a proper bugfix is needed in the long run, but that doesn't help us right now. K7L (talk) 17:31, 27 January 2013 (UTC)

Another thing to look at?[edit]

Most city or region articles should have a link right after the name. For example:

  • Xiamen (厦门; Xiàmén) [5] is a coastal city in Fujian Province in China.

The link should point to the local tourist board or local government.

I do not think a bot can fix these when they are either absent or pointing to the wrong thing. However, it should be straightforward to generate a list of articles without such links. Phrasebooks, travel topics, intineraries, talk pages and user pages can be ignored. Pashley (talk) 02:08, 7 February 2013 (UTC)

Using the bot to help crowdsource links to Wikipedia articles[edit]

I have been working for the last week to add interwiki links to Wikipedia articles, to test out the scope of the problem. As a result, I'm convinced that there are thousands of articles which could be linked, but are not. Linking these articles would help collaboration between Wikivoyage and other projects, as well as helping drive traffic from Wikipedia via adding links there to Wikivoyage articles which are sufficiently advanced to be useful.

Although I have linked a substantial number of articles, there are way too many articles like this for one person to finish linking within any reasonable time. However, I believe that this task can be handled quite easily by crowdsourcing the task.

Accordingly, I have put together an experimental tracking scheme for these articles, with the intention of helping this process. It uses Template:No Wikipedia link, which generates links to the hidden category Category:Articles without Wikipedia links. The template also takes an optional argument which may eventually be used to add articles to per-country subcategories of this maintenance category if needed. As an experiment, I've added this template to Brouage and Tierra del Fuego National Park.

I would now like to use my bot to add this template to several thousand articles, to start off the process. -- The Anome (talk) 21:11, 2 March 2013 (UTC)

  • Support from me. This type of approach worked nicely when we were migrating images from wts, so there's a precedent for using bots to tag articles for editorial purposes. -- Ryan • (talk) • 21:14, 2 March 2013 (UTC)
Update 1: see Wikipedia:Template:Coord missing for another example of using a bot-generated template and hidden categories to crowdsource manual effort to deal with a similar task. -- The Anome (talk) 21:18, 2 March 2013 (UTC)
Update 2: I've now scanned the most recent dump, and there appear to be over 6000 articles that are candidates for interwiki links to Wikipedia, but do not yet have them. That is both (a) quite a daunting backlog, and (b) a serious long-term opportunity to increase Wikivoyage's traffic visibility -- The Anome (talk) 23:46, 2 March 2013 (UTC)

Hi -- is anyone here able to give me a second endorsement for this, so I can go ahead with the edits? Thanks, -- The Anome (talk) 20:06, 4 March 2013 (UTC)

  • Support. In your opinion have we exhausted all opportunities for automation here? Have we done the interwiki thing to death (links to language versions linking to wikipedia language versions, linking to en.wp?) I'm not convinced that this is of a scale we can actually practically work with. --Inas (talk) 20:15, 4 March 2013 (UTC)
    • Not completely exhausted, no, but I've already handled most of the really low-hanging fruit via automated matching. Doing this will have two advantages: firstly, it will help coordinate the process of manual matching, (which experience suggests can be much more effective than you might think, as some people love to have tasklists they can work their way through -- for example, I worked my way through around 500 such articles in a month, almost one-tenth of the whole job), and secondly, it will hopefully let me see patterns in what remains, and in the patterns of matches being done by people, to provide inspiration for more fine-grained automatic matching processes. Experience from en: Wikipedia shows that this combination of multiple human and manual approaches is highly effective: the Wikipedia:WP:COORD project has worked its way through hundreds of thousands of articles to date in this way. -- The Anome (talk) 20:26, 4 March 2013 (UTC)
      More than happy for you to be right, of course. Let's see how we go. --Inas (talk) 20:28, 4 March 2013 (UTC)
  • Support. Links to Wikipedia articles are very needed, especially to prepare the ground for Wikidata, and also to one day benefit from the latitude/longitude info from Wikipedia. Nicolas1981 (talk) 06:18, 8 August 2013 (UTC)

Is this template still going to be used for something, or can we get rid of it? Texugo (talk) 18:07, 10 October 2013 (UTC)

  • Support, in principle at least. Where will the bot place the template in the article? Will it be in the position where the WP link would be placed, so that anyone who goes to add the WP link will see the template? Also, can a comment be added beside the template, saying "If you add a link to the Wikipedia article for this destination, please delete this template"? Nurg (talk) 07:35, 11 February 2014 (UTC)

User:Sumone's bot[edit]

I would like to run a bot using w:wp:AWB for various find-and-replace cleanup (such as fixing dmoz links). I have been using AWB on my regular account (in semi-automatic mode), but I would like to run it on automatic mode on a bot account. sumone10154(talk) 01:23, 12 December 2012 (UTC)

  • Support. Barring objection, as soon as one other support vote is registered I'll flip the bot flag on the account. -- Ryan • (talk) • 02:52, 12 December 2012 (UTC)
    In a gross abuse of power I've flipped the bot switch for this account, but in order to do things by the book it would still be good if one more person could add a "support" vote. -- Ryan • (talk) • 06:38, 12 December 2012 (UTC)
  • Support Can't wait to get all this things done and back to writing. jan (talk) 09:27, 12 December 2012 (UTC)
  • Support. The dmoz links really need fixing—the other language versions could also use this (at least the Dutch version). --Globe-trotter (talk) 20:12, 12 December 2012 (UTC)
Sure, I'll request bot approval on the other language versions as well. sumone10154(talk) 20:21, 12 December 2012 (UTC)

I've fixed about 2000 dmoz links; if there are still anymore broken ones, let me know and I'll see how they were missed.

I can also replace all wts:category links with links to commons categories. However, there are some categories that are named differently, or might not even exist on commons. My bot cannot check when this is the case; should I still do these replacements, or should they be done manually? sumone10154(talk) 02:01, 13 December 2012 (UTC)

I think if the wts: replacement cannot be done accurately then it would be better to skip that article, and we can either update it manually or wait until Sumone's bot 9000™ has been created to use secret technology to more accurately determine commons links (perhaps by following Wikipedia interwiki links and using the same Commons interwiki as the linked article?). -- Ryan • (talk) • 01:49, 14 December 2012 (UTC)
That last sounds interesting. I'll look into it. -- The Anome (talk) 20:03, 14 December 2012 (UTC)
Update: To get a direct interwiki link to a commons category such as [[Commons:Category:London]], which seems to be what is required here, I think the easiest path to follow would be Wikivoyage article -[interwiki]-> Wikipedia article -[category link]-> Wikipedia category -[interwiki]-> commons category. There's also Template:Commons_category, on enwiki, but it's not present on many pages, for example, enwp:London, where it might normally be expected. Coupled with all the necessary cross-checks, this seems like a lot of work for the benefit involved: so, all things being considered, I'll give this one a pass for now, and look for some bot tasks with greater cost-benefit ratio in the short term. -- The Anome (talk) 17:53, 16 December 2012 (UTC)
We could also remove all wts links first and then add Commons links en masse. --Globe-trotter (talk) 20:55, 14 December 2012 (UTC)
I've now added Commons links en masse, and am looking to remove the wts: links from all articles with Commons links, prior to surveying the remaining wts: links in other articles -- please see the discussion in the sub-section above. -- The Anome (talk) 19:19, 9 January 2013 (UTC)
All Yes Done. See above. -- The Anome (talk) 21:04, 15 January 2013 (UTC)

Crochet.david.bot[edit]

Crochet.david (talk) 20:29, 4 January 2013 (UTC)

  • Support. -- Ryan • (talk) • 20:42, 4 January 2013 (UTC)
  • Support and now flagged as bot. --Peter Talk 21:07, 4 January 2013 (UTC)


CarsracBot[edit]

  • Botmaster: User:Carsrac
  • Bot's name: User:CarsracBot
  • List of bot flags on other projects: full list (all wikipediaprojects) Globalbot bit.
  • Purpose: interwiki
  • Technical details: pywikipediabot, latest versions

Carsrac (talk) 17:30, 18 January 2013 (UTC)

  • Support K7L (talk) 17:44, 18 January 2013 (UTC)
  • Comment I'm in favor of automating as many tasks as we can possibly automate, but could you clarify what this bot does? Does it just search for articles with the same name, does it ensure that interwiki links to a target article are also present from a target article? I'd like to better understand what this does before flipping the bit. -- Ryan • (talk) • 18:11, 18 January 2013 (UTC)
    • It does check the interwiki links and adds and removes the links. It does actively seek interwiki links by searching for articles with the same name or translation of the that name. With places and other geographical names and the job is very simple. Other wiki's are much harder to interwiki for a bot :) Carsrac (talk) 19:56, 19 January 2013 (UTC)
  • Comment This looks to be interwiki.py, a script in the pywikipediabot bundle which has been used for a decade on Wikipedia to generate interwiki links. If en:Germany links to fr:Allemagne and fr:Allemagne links to de:Deutschland, the script creates the "in other languages" links to link the en: and de: articles to each other, as well as creating corresponding reverse links (so fr:Allemagne and de:Deutschland link to en:Germany). It is bright enough to detect if an existing "in other languages" link points to a disambiguation, a redirect or a deleted page. K7L (talk) 18:45, 18 January 2013 (UTC)
    • Sounds good then. Support from me. -- Ryan • (talk) • 18:56, 18 January 2013 (UTC)
  • Support and flagged. --Peter Talk 19:30, 18 January 2013 (UTC)
      • The disambig detection does not fully works on all the wikivoyageproject at a way that the pywikipediabot's can pick it up automaticly. I'm aware of this problem I will at this moment not interwiki disambig pages until it's is solved. Carsrac (talk) 21:16, 24 January 2013 (UTC)

RileyBot[edit]

Riley Huntley (talk) 09:40, 19 January 2013 (UTC)

  • Support. Since we don't get as many edits as Wikipedia then increasing the threshold to 6 hours or even one day would be fine, but assuming that cleaning the sandbox is standard practice on Wikimedia wikis then we might as well do it here, too. -- Ryan • (talk) • 20:00, 19 January 2013 (UTC)
  • Support and flagged. I think 6 hours would work well. --Peter Talk 09:04, 20 January 2013 (UTC)
I am boldy plunging forward and doing the same for Template:Graffiti wall. -- Cheers, Riley 21:18, 20 January 2013 (UTC)

RileyBot 2[edit]

-- Cheers, Riley 09:46, 26 January 2013 (UTC)

  • Support. -- Ryan • (talk) • 16:21, 26 January 2013 (UTC)
  • Support useful function. --Rschen7754 04:56, 27 January 2013 (UTC)
  • Support Nice and simple Addshore (talk) 23:11, 10 February 2013 (UTC)

User:Hoo User Page Bot[edit]

This script was blocked for lacking a nomination, so I'm nominating it here. From the description:

This bot syncs user and user talk pages across all wikis. Syncs can be requested here. Feel free to complain if it does anything it shouldn't. Every request gets performed and reviewed by Hoo man.

This looks to be a standard bot run across Wikimedia projects, so I see no harm in letting it run here.

  • Support. -- Ryan • (talk) • 05:14, 2 February 2013 (UTC)
  • This falls in the category of globally-used bots (not the same thing as global bots), and is essential for global users like myself who need userpages on all WMF wikis (over 700+). User:EdwardsBot is another example, to provide messaging across all wikis, and has already left messages here. This bot should be allowed to operate here, out of goodwill for other projects. --Rschen7754 06:16, 2 February 2013 (UTC)
Thanks for flagging the bot. It will keep editing this wiki then. - Hoo man (talk) 14:29, 3 February 2013 (UTC)


User:inasbot[edit]

  • Botmaster: user:Inas
  • Bot's name: User:Inasbot
  • Purpose: Rename Contact Header to Connect in main namespace travel guides.
  • Technical details: pywikipediabot, replace.py

Since, this needs to be done, and noone else has volunteered to run it.. --Inas (talk) 10:22, 28 February 2013 (UTC)

  • Support - Thanks for picking up the slack. Have you successfully tested it yet? JamesA >talk 10:27, 28 February 2013 (UTC)
Yep. I've run over a heap of files in the main namespace, so I'm confident that part will be fine. Whether I can actually sweep the entire main namespace without cluttering up recent changes with bot edits for months is another matter, but we'll see. I've set up the testcases at User:Inasbot/testcases. If there are any others you feel it needs to pass/fail, you can add them. --Inas (talk) 11:15, 28 February 2013 (UTC)
  • Support but please create a page so people know what the bot is for :) --Rschen7754 10:55, 28 February 2013 (UTC)
Done. --Inas (talk) 11:17, 28 February 2013 (UTC)
  • Support. I've flagged this account as a bot. -- Ryan • (talk) • 15:52, 28 February 2013 (UTC)

User:Traveler100bot[edit]

Removal of unused templates appears to have general support.--Traveler100bot (talk) 14:02, 3 March 2013 (UTC)

  • Support. There are far too many of these to be fixable manually. K7L (talk) 18:09, 3 March 2013 (UTC)
  • Support. -- Ryan • (talk) • 19:13, 3 March 2013 (UTC)
  • Limited Support. As per discussion in pub. However, if I understand the current method, it would remove cityguide, etc al, from articles with that are stubs or unranked. This means we lose the fact that they are cities, or countries, etc. I'd suggest we need a stubcity, unrankedcity, etc so this process doesn't involve information loss. --Inas (talk) 21:40, 3 March 2013 (UTC)
    • Easiest way to avoid that would to be to run the bot just against articles which are outline/usable/guide/star? K7L (talk) 21:50, 3 March 2013 (UTC)
      I noticed there were only 300 or so unranked articles, so I ranked them. That should get rid of most of the problems. I don't think the stubs are so much of a concern. --Inas (talk) 05:56, 4 March 2013 (UTC)
      • Nice job, solves one difficulty. Have just run a test run. If no one sees any issues will do the rest later in the day.--Traveler100 (talk) 06:45, 4 March 2013 (UTC)
        Spelling mistake in the summary. --Inas (talk) 09:54, 4 March 2013 (UTC)
  • Account now has the bot flag. -- Ryan • (talk) • 21:43, 3 March 2013 (UTC)

User:Traveler100bot expansion[edit]

I assume I need to apply or at lease check again if I expand the logic? Would like to do the following:

Article with {{outline}} change to {{outlineregion}} if section title ==Cities== exists on the page.

Have check with a few manual checks of semi-automatic run and appears to work fine.--Traveler100 (talk) 15:39, 9 March 2013 (UTC)

  • Support. --Peter Talk 15:53, 9 March 2013 (UTC)
  • Support. Makes sense. -- Ryan • (talk) • 16:39, 9 March 2013 (UTC)
  • I have not been following these changes in any detail. Why should a city be labelled as a region, rather than some tag like "outlinecity"? Pashley (talk) 16:54, 9 March 2013 (UTC)
    • It is that some are labelled with just "outline" or "usable" , could be city, region, park or country. Check out current status at Wikivoyage:Article status stats. Also I have found and corrected quite a few already that are marked as both city and region or region and park. --Traveler100 (talk) 18:22, 9 March 2013 (UTC)

User:Traveler100bot expansion 2[edit]

Change articles with {{outline}} to {{outlinecity}} or {{outlineregion}} or {{outlinepark}} based on following logic-

  • if contains ==Cities==|==Towns and villages==|==Regions==|==Cities and towns==|is a region
    • make outlineregion
  • if contains is a city|is a town|is a village|is the main city|is the capital city
    • make outlinecity
  • if contains ==Fees/Permits==
    • make outlinepark

Have tested manually on 120 pages edited 60 other skipped. --Traveler100 (talk) 13:48, 29 March 2013 (UTC)

  • Support. Sounds simple enough - go for it. -- Ryan • (talk) • 14:56, 29 March 2013 (UTC)
  • Support. --Peter Talk 17:51, 29 March 2013 (UTC)

User:Traveler100bot expansion 3[edit]

Probably the last possibility to do anything automatic on the deprecated tags, and has a small risk of making a mistake.

  • Change articles with {{outline}}
  • if contains == Sleep ==
    • make outlinecity
  • I'm not sure that will work - there are a fair number of regions and countries that have a "Sleep" section. See USA#Sleep, Finger Lakes#Sleep, etc. It might be best to just tag the remaining articles and try to fix them manually. -- Ryan • (talk) • 15:57, 3 April 2013 (UTC)
  • I am working on the principle that I have already removed all the articles with a section called Regions, Cities, Villages, Towns or Fees and variations on that theme, so what is left (2351) is most probably note a region or country. --Traveler100 (talk) 18:04, 3 April 2013 (UTC)
    • Could you run your bot on perhaps 100 articles, and we'll then do a quick review? Based on your most recent comment it sounds like this should be fine, but it would be good to ensure we don't get too many incorrect tags before updating 2351 articles. -- Ryan • (talk) • 18:45, 3 April 2013 (UTC)
      • Done first 100 in Category:Outlines, shown as contributions made on 6 April. --Traveler100 (talk) 07:22, 6 April 2013 (UTC)
        • I did a spot check and most look correct. Altar Desert appear to be a region, but it's using the city template so tagging it as such doesn't make things any worse than they are now. I'd say go ahead with the rest. -- Ryan • (talk) • 18:11, 6 April 2013 (UTC)

User:Traveler100bot task 4[edit]

After running a number of variations with manual check of each edit for a few days I now think I have a script that will correct telephone numbers. It does not correct all and will skip many that have odd formats. So this will not totally remove, but will hopefully reduce the numbers of Listing with phone missing country code and Listing with phone format issue pages. This is not just to have some consistency but also to allow dialling from links in Wikivoyage on a smart phone. Would now like to propose running it as a bot.Traveler100 (talk) 20:00, 1 October 2013 (UTC)

Sure, go ahead. Ikan Kekek (talk) 23:12, 1 October 2013 (UTC)
  • Can you provide some links to sample edits that the bot will make? I'm fully in support of cleaning up some of the messy phone numbers on the site, but I'd like to better understand what exactly is going to be changed. -- Ryan • (talk) • 01:58, 2 October 2013 (UTC)
edits with current logic --Traveler100bot (talk) 05:33, 2 October 2013 (UTC)
Here, the bot removed hyphens; was this based on the bot's knowledge that 7-digit-dialing is allowed in the 518 area code? Or was it a general policy to remove hyphens between the area code and the local number? LtPowers (talk) 16:18, 2 October 2013 (UTC)
It removes hyphen been area code and local. Which is my understanding of current recommendation. --Traveler100 (talk) 18:54, 2 October 2013 (UTC)
Based on the edit history, is the bot just doing US phone numbers? When I was doing the text-to-template conversions for listings I found a wide variety of formats used for non-US listings. Assuming this is only for the US, and assuming LtPowers' concerns can be addressed, then support from me. -- Ryan • (talk) • 00:09, 3 October 2013 (UTC)
No I intend to do it for each country, just the easiest way to ensure getting the country code correct is to do one country at a time (have already done Germany manually). Do I assume from the comment that the format <country code>blank<region code>blank<local number with dashes> should only be applied to USA? --Traveler100 (talk) 05:30, 3 October 2013 (UTC)
Be careful with other countries - I encountered all manner of strange formats, particularly with some of the India articles, and my parser often failed to even recognize the numbers as phone numbers. As to formatting specifics, hopefully someone else can answer that - I've never cared enough about the nuances of date formatting, phone number formatting, etc and have left it to others to battle over those details. -- Ryan • (talk) • 05:39, 3 October 2013 (UTC)
Traditionally we have used hyphens to indicate what parts of a phone number can be dialed locally. So, in the U.S., we would use "+1-800-555-5555" for toll-free numbers, "+1 212-555-5555" for areas with mandatory 10-digit dialing, and "+1 518 555-5555" for areas with 7-digit dialing. LtPowers (talk) 14:29, 3 October 2013 (UTC)
Actually that should be "+1-212-555-5555", all eleven digits required on a local call. K7L (talk) 02:28, 4 October 2013 (UTC)
Then why do they call it ten-digit dialing? =) LtPowers (talk) 12:17, 4 October 2013 (UTC)
They don't. From [6], ten-digit dial is a Boston-style number " 617 MA 10D" but NYC/LA/Chicago are "1+10D". Effectively, there is no such thing as a flat rate local call in these three cities or their immediate area, everything is metered. A call across the street to "PEnnsylvania 6-5000" from Madison Square Gardens is +1-212-736-5000, eleven digits, same as if the call were made from Honolulu. "212 NY 1+10D" K7L (talk) 18:14, 4 October 2013 (UTC)
Ah, sorry; I didn't realize you meant specifically the 212 area code. In that case, I erred in my choice of example. I was just trying to think of an overlaid area code off the top of my head, but I inadvertently chose one with special rules attached. Feel free to substitute some other overlaid area code like 416, perhaps. LtPowers (talk) 21:04, 4 October 2013 (UTC)

User:SteveRBot[edit]

  • Botmaster: user:SteveR
  • Bot's name: User:SteveRBot
  • Purpose: Ukrainian wikivoyage was launched recently. Bot adds interwiki from uk.wikivoyage and other voyages if needed.
  • Technical details: Bot uses pywikipediabot library. On en.wikivoyage bot will use interwiki.py only. You can read about it here. --SteveR (talk) 17:25, 1 April 2013 (UTC)
  • Support. -- Ryan • (talk) • 17:41, 1 April 2013 (UTC)
  • Support. Pashley (talk) 18:26, 1 April 2013 (UTC)
  • Flagged. -- Ryan • (talk) • 18:40, 1 April 2013 (UTC)

User:Traveler100bot pagebanner[edit]

Add pagebanner template to location pages in Austria as describe at User:Traveler100bot and discussion at Wikivoyage_talk:TOC/Banner. --Traveler100 (talk) 05:12, 14 April 2013 (UTC)

  • Support #1 JamesA >talk 06:56, 14 April 2013 (UTC)
  • Support -Shaundd (talk) 07:38, 14 April 2013 (UTC)
  • Support - I usually stay out of bot discussions, but the reasons for this bot are clear. Ikan Kekek (talk) 13:34, 14 April 2013 (UTC)

User:Kolega2357-Bot[edit]

--Kolega2357 (talk) 23:08, 12 May 2013 (UTC)

What is your problem Rschen7754? You are so complexed from me. Bot has its own folder, not the other Wiki projects in the same folder. Do not make war changes. Run it like this python interwiki.py -autonomous -start:A -lang:en. --Kolega2357 (talk) 07:45, 13 May 2013 (UTC)

  • Note: I've blocked the bot as it already started running, without approval. --Rschen7754 08:00, 13 May 2013 (UTC)
    • Further note: this bot operator has gone to several wikis, filed a request similar to the one above, and then started running his bot immediately, with no flag or local approval. I am trying to follow up on this. --Rschen7754 08:22, 13 May 2013 (UTC)

@Pashley: I placed request everywhere a local bureaucrats. --Kolega2357 (talk) 09:20, 16 May 2013 (UTC)

Can someone administrator to unblock my bot has been more than a month he is currently inactive. --Kolega2357 (talk) 21:47, 1 July 2013 (UTC)

Why? There's no reason that it needs an unblock. --Rschen7754 22:05, 1 July 2013 (UTC)

There a reason I saw this and Wikivoyages migrating interwiki to Wikidata. I'm here to signatured. --Kolega2357 (talk) 22:13, 1 July 2013 (UTC)

That does not mean that we have to give you bot access; there's plenty of other bot operators. --Rschen7754 23:12, 1 July 2013 (UTC)
Oppose for the time being. I can't even figure out what the proposed task of this bot is. We haven't started using Wikidata yet, and if the proposal is just updating interwiki links, we already have working and tested bots handling those; I don't think we'd need or want multiple bots trying to do the same task simultaneously. If the proposal is something else, that needs to be explained, covering the three key points listed at the top of the page. -- D. Guillaume (talk) 23:16, 1 July 2013 (UTC)

Why so biased user Rschen7754? I'm want to get a bot flag here. If all interested operators bot can do it in the near future, then why can not I? What is the purpose of it all can, and I can not? Assume good faith. --Kolega2357 (talk) 23:27, 1 July 2013 (UTC)

Approval for a bot flag here requires that there are no outstanding objections to the request, and since current objections are related to the concerns raised at commons:Commons:Administrators' noticeboard/Archive 41#Kolega2357 you will need to provide a link to a discussion or some other evidence demonstrating that those issues have been successfully resolved. Until that is done, it is probably pointless to continue to request permissions here. -- Ryan • (talk) • 23:53, 1 July 2013 (UTC)

No more such problem I no more FileMover rights on Commons. I have a bot flag on several Wikivoyage projects. --Kolega2357 (talk) 00:21, 2 July 2013 (UTC)

Yes, because they are mostly inactive and/or were not aware of your crosswiki record. --Rschen7754 02:11, 2 July 2013 (UTC)

User:Wrh2Bot[edit]

XML listings to templates[edit]

Designed to convert the legacy XML listings to templated listings per Wikivoyage talk:Listings#Tags to templates. See Special:Contributions/Wrh2Bot for a sample run against the first 100 articles (alphabetically) in the main namespace, and User:Wrh2Bot/ListingsToTemplate for log output. The bot converts well-formatted XML to a template. In cases where the bot cannot reliably convert the listing (due to mis-matched tags or other errors) then the bot skips the listing and writes a message to the log (I have to manually upload logs here). If approved I can probably kick this off over the weekend when I'm back home and will have time to monitor it. -- Ryan • (talk) • 00:19, 16 May 2013 (UTC)

  • Support. --Peter Talk 04:22, 16 May 2013 (UTC)

External links[edit]

I've got code ready that will convert links of the form '''Bold text''' [http://www.example.com] to '''[http://www.example.com Bold text]''', per the latest changes to Wikivoyage:External links. This code currently works ONLY for the specific pattern mentioned - if the text preceding the link is not bold it won't match, but this is a start that should get a significant portion of the site updated. See the following tests:

I'll investigate options for converting additional links in a future update, but it's a tricky job. -- Ryan • (talk) • 04:35, 2 July 2013 (UTC)

  • Support. Those tests were successful, and I can't think of an instance where this wouldn't be desirable. Tackling this chore piecemeal like seems a good idea. It probably goes without saying, but this is just for the Mainspace, right? --Peter Talk 05:00, 2 July 2013 (UTC)
    Yes, mainspace only. If it gets the OK I'll run it in a few small batches so I can keep an eye on things, but I don't expect there will be issues with a pattern this simple. While this bot won't convert all footnote links, it should at least get the official city link at the start of articles, which is a prominent enough link to make this worth running. -- Ryan • (talk) • 05:06, 2 July 2013 (UTC)
  • Support. Definitely a job for bots, and a cautious pattern like that should avoid false positives well. Agreed that it needs to remain mainspace-only for now, as your sandbox test pages in User: space are clearly still excluded from the new CSS. -- D. Guillaume (talk) 05:07, 2 July 2013 (UTC)

Done. Just over 7000 articles updated. -- Ryan • (talk) • 02:30, 3 July 2013 (UTC)

Nice. Are there any in the pattern ''Italic text'' [http://www.example.com]? If so, that should be an easy change to your code, so straightforward to do next. Pashley (talk) 04:45, 3 July 2013 (UTC)
Peter started a list of potential patterns at Wikivoyage talk:External links#Bot suggestions, although we'll need to be careful as it is very easy to pick up false positives. The italics pattern would be a good one to add to the list, and it definitely shouldn't trigger any false positives. -- Ryan • (talk) • 04:56, 3 July 2013 (UTC)

Text-to-template Conversion[edit]

For a couple of weeks I've been testing a bot that converts plain text listings to use the appropriate version of Template:Listing. More details can be found at Wikivoyage talk:Listings#Text-to-template conversion, but the high-level overview is that the bot is pretty conservative, and will only convert fields when it has a very high probability of getting the conversion right. If it is unsure, the listing is either not converted, or if a particular field is questionable it is simply left as part of the listing "content" field where someone can manually move it to the right field later on. There is still the rare false positive - for example, in this edit the opening sentence "This 3 small churches built around St.Bogorodica Perivlepta" was interpreted as containing a street address ("3 small churches built around St.") - but these are pretty rare and easily cleaned up.

We'll never get a bot that can convert with 100% accuracy (or a human who won't make occasional mistakes, for that matter), and after literally testing and reviewing hundreds of edits I think this bot is worth running, but will wait a few days for any objections (and the required support votes) before doing so. See Special:Contributions/Wrh2Bot for a history of bot testing - to this point I have manually reviewed every single edit made and tweaked it to improve accuracy, but would now like to unleash the bot to run on its own. -- Ryan • (talk) • 04:59, 19 July 2013 (UTC)

  • Support. While it may make a little mess on some of the items, all of the items it's converting are basically a mess already. It's too big a job to do by hand; this will save us a lot of time in the long run. --Peter Talk 05:16, 19 July 2013 (UTC)
  • Support - This will further modernise and standardise our guides and ensure that we are more ready for future features such as dynamic maps. Thanks for your hard work on this! James Atalk 06:33, 19 July 2013 (UTC)
  • Comment. The bot is done through Rye (New York) but I need to head out for the evening so I'm going to halt it and will let it finish when I'm back to monitor things again. -- Ryan • (talk) • 00:31, 21 July 2013 (UTC)
    The bot has completed with 5687 articles updated. -- Ryan • (talk) • 19:17, 21 July 2013 (UTC)

User:Addbot removing interwiki links[edit]

See meta:User:Addbot. Wikivoyage will be enabled on wikidata next week. My bot taking interwiki links from pages, puts them on wikidata and then removes interwiki links that are no longer needed from articles. This is a task that is currently running on all language wikipedia projects with over 14 million edits and it would be great to also get approval here! Addshore (talk) 12:22, 19 July 2013 (UTC)

  • Question - is this a task that global bots are automatically approved to do? If so then global bots can run here without additional approval. If this isn't a global bot task, then support from me for your specific bot. -- Ryan • (talk) • 14:52, 19 July 2013 (UTC)
  • Support trusted user and trusted bot, though it may be moot with the global bot issue (I'm not entirely sure on that myself). --Rschen7754 23:44, 19 July 2013 (UTC)
  • Indeed the task is within the scope of global bots, hence the global bot flag :) Addshore (talk) 07:28, 20 July 2013 (UTC)

User:Andyrom bot removal of unknown Print= parameter[edit]

I've seen that several images have this parameter that has been imported from WT. Both in voy and WT this parameter it's not managed by the parser, so it must be removed. Although it's an easy task I'm making right now 50 test modification so anyone can verify them. --Andyrom75 (talk) 18:44, 5 September 2013 (UTC)

It was handled by Wikitravel Press's parser. There's probably no reason to keep them, unlike other WTP legacy syntax, but neither is there any urgency to remove them. LtPowers (talk) 13:43, 6 September 2013 (UTC)
Peter, thanks for giving me that background, I didn't know anything about it. And I didn't know that you were selling book! :-D If you know other syntax (print= a part), just let me know and I'll remove them as well. Clearly it's not urgent, but I tend to eliminate what is useless to avoid that it could create confusion within the editors or in the layout like in those cases where the images has no description, so the caption is "print=...." --Andyrom75 (talk) 18:56, 6 September 2013 (UTC)
Peter who? =) LtPowers (talk) 18:32, 7 September 2013 (UTC)
  • Support. Might as well clean that up. -- Ryan • (talk) • 14:22, 6 September 2013 (UTC)
  • Support. Seems reasonable. Based on existing test contributions, I am happy with allowing this bot at this time. --Saqib (talk) 06:22, 7 September 2013 (UTC)
  • Support. This parameter has been puzzling me for a long while, so it's nice to get an authoritative view from LtPowers that it is now obsolete and redundant. They should be removed because, like some HTML, they may baffle newbies (and oldies like me!) --W. Frankemailtalk 13:28, 7 September 2013 (UTC)
  • Flagged. -- Ryan • (talk) • 22:48, 7 September 2013 (UTC)


This bot approval applied only to removal of print parameters from image tags. It did not have authorization to remove other items. LtPowers (talk) 19:01, 9 September 2013 (UTC)

I've cleaned all the WTP legacy syntax explained in the page you linked. Going into details:
  • Removing of all the print= occurrences
  • Removing of all the angle= occurrences
  • Removing of the redundant text between PRINT "tag-comment" (redundant because all these occurrences have the same/similar text inthe article)
  • Removing of the WEB-START & WEB-END "tag-comment" leaving the text between them
If you think I've made a mistake on that let me know and I'll revert my work without any problem. --Andyrom75 (talk) 21:44, 9 September 2013 (UTC)
To make a perfect job, I suggest to orphanize also the Template:Pagebreak and Template:Index, both used by WTP and applied on few articles. Feel free to discuss about it internally. I don't think I'll be strictly necessary for this operation, but let me know if I can help on other topics. --Andyrom75 (talk) 21:50, 9 September 2013 (UTC)
I apologize if I wasn't clear. Your bot was authorized only to remove "print=" attributes within File tags. None of the other stuff you mentioned was even brought up in this discussion, and your bot had no authorization to remove them. The "angle=" attribute removal is not harmful, but removing the PRINT and WEB-START/END tags was potentially so. Each of those cases needs to be addressed individually, and a bot should not be removing them wholesale. Period. Please revert those changes ASAP. LtPowers (talk) 23:05, 9 September 2013 (UTC)
If you think it's critical & urgent the quickest way is that an admin revert all the changes of the bot with just one click. Otherwise it will take time, and although WTS doesn't work neither here and in WT, I don't want to slow down your request. --Andyrom75 (talk) 23:14, 9 September 2013 (UTC)
I didn't rollback all of the edits because I didn't want to lose the good ones. It's probably not urgent; it's just that the longer we go before undoing them, the harder they may be to undo. Make sense? LtPowers (talk) 01:49, 10 September 2013 (UTC)
If it's not so urgent maybe could worth to take few minute to understand if all can be consider "good ones" and if we really want to restore the other tags that are not used by the wikimedia parser. For example, "angle=" was used jointly with "print=" so why to keep the first while we are removing the second?
Inside the PRINT tag comment, you can find two kind of commented wikitext.
  • The same image already shown in the article (usually few lines above)
  • The same text already written in the article (I remember an infobox that used a different template, and the regionlist description without the use of the regionlist template)
So, being redundant and not being shown in any media (because it's commented and not managed differently) why to keep it?
The WEB-START/END tag were used to delimit a text shown only via web, but currently this TAG doesn't work and the text are shown in any media, and I would say correctly, like the ones that are in a phrasebook that contains almost the whole the article. As written above, why to keep a useless comment?
PS Sorry for the late answer but here was sleeping time :-) --Andyrom75 (talk) 07:12, 10 September 2013 (UTC)
As I said above, "Each of those cases needs to be addressed individually, and a bot should not be removing them wholesale." Are you certain that every case is the same? LtPowers (talk) 17:06, 10 September 2013 (UTC)
If you feel that we have enough time we can discuss about them one by one (I don't think it would take so much time). I'm not sure to understand what do you mean with "every case is the same"; could you rephrase it? --Andyrom75 (talk) 19:28, 10 September 2013 (UTC)
You seemed confident that in every instance in which these tags were used, they could be removed without concern. I was asking if you were sure about that. Anyway, I was kind of hoping that some other contributors would chime in here. LtPowers (talk) 02:06, 11 September 2013 (UTC)
Aren't all the tags just remnants of WT Press? To determine what made the print guide and what was web only? I can't see they have any use, and although I agree with the point that they shouldn't have been removed without discussion, I think we have a short window to discuss to see if we should keep before reversion. --Inas (talk) 04:03, 11 September 2013 (UTC)
Thank LT for the clarification. I'm pretty sure that the tag can be removed (otherwise I wouldn't have done it). And I haven't processed them all in an automatic way. I've made test to be sure that the layout wouldn't have any kind of change, and I've verified, expecially for the PRINT tag the presence of the equivalent removed text (inside the tags) outside those tags. The WEB tags should be used to hidden the text inside those tags so it has 0 impact on screen, and without a dedicated parser (like WTS) that process them in that way, those tag are useless. Furthermore, in the Japanese prhasebook the WEB tags were used in a strange way because they included almosto the whole article, and it doesn't make sense to avoid its print on paper. --Andyrom75 (talk) 10:19, 11 September 2013 (UTC)
Unfortunately, yes, tags are not currently functional, as far as preventing web-only content from being printed. (Jani deleted Template:Web and all of its uses a while back, even though we could have maintained its functionality by including it in Category:Exclude in print. The WTP-only method of using HTML comments doesn't have this advantage.) I would just hate to lose this semantic information -- even if it doesn't actually do anything at the moment -- the way we lost Template:Web. LtPowers (talk) 13:22, 11 September 2013 (UTC)
This functionality could be easily restored in a different way: through CSS, where you can classify the content for specific devices like mobiles & printers, or to avoid that those devices would receive specific content. A parser it's not necessary for this task. If interersted to go deeper on this topic, maybe we can keep discussing in another page. --Andyrom75 (talk) 15:56, 11 September 2013 (UTC)
Yes, there are ways to replicate the functionality, but some of it may not be needed anymore if current printing methods work better than WTP's did. That's why I said we have to go through each case individually rather than just remove them all. LtPowers (talk) 16:51, 11 September 2013 (UTC)
The fact is that the logic on how CSS works is different, so different is their implementation and cannot be (for example) replaced 1 by 1. Usually the CSS should be included in template so that you'll have the control over the whole site instead of the single piece of article. Just to mention an example. I've removed a bunch of text that was a "printable copy" of a regionlist. The right way to do it is to work on Template:Regionlist and not on a specific article. --Andyrom75 (talk) 17:27, 11 September 2013 (UTC)
Granted, but that seems all the more reason to go through one-by-one instead of simply removing them. LtPowers (talk) 00:17, 12 September 2013 (UTC)

I'm still inclined to think that the basis for them was editorial by WTP. Personally, if I get a print, I'd like to have all the info that is present on the web with automatic formatting applied as appropriate. I don't want anyone making a manual call on what information should be excluded from my print. The fact that they aren't even currently functional, is the clincher. --Inas (talk) 01:43, 12 September 2013 (UTC)

Inas, I agree with you, also because, especially for printing, doesn't comes up to my mind any site that use properly (cum grano salis ;-)) a customized printed version. A good and complete article should be passed as it is to all the media. However I do not exclude that with a deep study, few templates could be modified to show differently the same content (NOT a different content! just the layout). --Andyrom75 (talk) 09:49, 12 September 2013 (UTC)

Fine, if we want to do this the hard way. Here is an example where Template:Web was used to exclude notations that make sense only in a linked environment. Here is another case where it was used to exclude a disambiguation notice, which is completely useless in print. Here (see last change) is a case where it was used to avoid printing a lengthy URL that is very similar to (and accessible from) a link earlier in the paragraph. Here it was used to avoid printing a link to a PDF. And that's just Template:Web; I haven't even gotten to the WEB-START/END cases. LtPowers (talk) 14:08, 12 September 2013 (UTC)

I think the way was used that template it's not so much different from how the WTS works. The principle is the same, with just the technical benefit that it's easier to manage, so I wouldn't resurrect it. If you are interested to differentiate the output, you should think different. Let's talk of one of the examples that you have mentioned: the disambiguation note. If I'm not wrong we should have a template for that, maybe is in en:voy as well. So the idea is to use that template any time and everywhere you need to show a disambiguation note, and then to apply the CSS patch to that template. Otherwise you'll flood all the article with that too general template that only complicate the editor's lives. IMHO this kind of patches, to be effective, should be transparent, otherwise people would naturally wreck this kind of initiatives just not applying them. --Andyrom75 (talk) 16:40, 12 September 2013 (UTC)
What do you mean by "WTS"? Also, I'm not sure we disagree; I'm all for putting this stuff into templates. My concern is that we are losing this semantic information by removing the markup without supplying a replacement, even if the replacement is currently not fully functional. LtPowers (talk) 17:21, 12 September 2013 (UTC)
Sorry I've just mistype, I meant WTP :-) In my opinion we are not losing anything valuable, because the cleaned tag are not useful for the scope. I suppose (and I hope) that WTP was designed before the CSS was able to managed the different media layout, because nowadays, no one would design it in that way, it's too complicated from both programming and using point of view. --Andyrom75 (talk) 20:12, 12 September 2013 (UTC)
The point is, I don't want to miss out on the link to the Chicago Skyline Guide just because I print something. When WTP were publishing stuff they did. They wanted a published book without these hooks. We don't want to exclude this information from the print version any longer. --Inas (talk) 20:51, 12 September 2013 (UTC)
Why not? What good is a disambiguation notice in a printout? LtPowers (talk) 23:15, 12 September 2013 (UTC)
We're not addressing the issue. These things were put in the articles for books published by WTP. If we don't want otheruses to render in certain views, then we should make that change to the template/css. We're not preserving anything here by keeping these formatting guidelines for WTP in the articles. Keeping these won't stop otheruses from appearing in the 99.9% of print article in which WTP never got involved. --Inas (talk) 23:48, 12 September 2013 (UTC)
I realize this; what I'm failing to communicate is that keeping these artifacts allows us to easily find and fix cases where we might want or need to maintain that functionality. If they are removed, they become that much harder to track down and replace. LtPowers (talk) 01:13, 13 September 2013 (UTC)
I just think the chances of us ever actually manually marking up our articles for a print view is so wildly remote, that I think having this info is still in the history is more than sufficient. It isn't as if someone is going to add any more of these, or we are going to use the same tags. The target articles can always be identified. However, if we go down this path, the only way I can ever see this working is templates like otheruses could be excluded from a print view, and these tags don't help us there. I understand it can be difficult to let go of the WTP idea, but I don't see it ever being reborn in that form. Any future view is going to be some form of automation applied to existing templates and css, and not manual classification. --Inas (talk) 01:29, 13 September 2013 (UTC)
I don't see why; there's no reason we can't write guides for print as well as for web use. LtPowers (talk) 02:04, 13 September 2013 (UTC)
Lt, IMHO you should separate the two discussions.
  1. WTP implementation is totaly wrong, and can't be an example for anything right now. This discussion should close this thread.
  2. Having a different layout for other media (printer included) is possible but must be rethink from scratch. This discussion should be held in a dedicated thread. If you think that it's a strategic topic, you can open an expedition (if it doesn't already exist ... I hanve't checked yet).
IMHO too also Template:web is a wrong approach, but this topic is related to the second point and not to the first one. --Andyrom75 (talk) 06:53, 13 September 2013 (UTC)
I'm not the one conflating the two discussions. The point remains that your bot exceeded its approved remit. We should revert the excess changes and do them the right way, by replacing them where necessary with functional code, not simply removing it and forcing us in the future to dig through histories to find them. LtPowers (talk) 13:18, 13 September 2013 (UTC)
As you prefer, in fact my first replay was you can revert them all, but in my opinion this would reintroduce useless code in en:voy. Believe me, I have no personal interest on convincing you (a funny quote says that opinions are like testicles, everyone has their own :-D ), I've just tried to explain you which would be the right solution to the different media issue. --Andyrom75 (talk) 13:49, 13 September 2013 (UTC)