Jump to content


From Meta, a Wikimedia project coordination wiki
Meta-Wiki Importer icon
Meta-Wiki Importer icon

Thaur's twa types o import, baith can be foond thro Byordinar:Import:

  • transwiki import, cried interwiki import aes weel: import pages strecht fae anither wiki; the settins o the destination wiki determine whit soorce wikis ar enabled; message with id 'import-interwiki-text' (talk) kiths; efter "Transfer pages into namespace" ye can speceefie ae tairget namespace; the optie "aw" actuallie means "the same aes the oreeginal".
  • uplaid import: import ae file in ae byordinar XML format produced bi exportin pages fae anither wiki; message with id 'importtext' (talk) kiths;

See the page Importer fer information aneat the uiser groops performin thir actions.

Transwiki import

Oan monie Wikimedia wikis the transwiki import is disabl't aes weel, it gies message with id 'importnosources' (talk): "No wikis from which to import have been defined and direct history uploads are disabled." Houever, pages fae commons:, foundation:, w:, cs: n fr: can be importit til Meta the nou, n pages fae Meta can be importit til mw:. The act o importin is eikit til the page histerie n til the log/import.

Gif aen importit page haes the same name aes aen exeestin page in the tairget wiki, the page is wruttenower gif the imported page is newer (accordin til the timestamps). Gif ae mistak happent durin the import than ye micht fynd the import is pairtiallie compleate (some pages imported, bit na aw). Sin pages ar wruttenower, ettlin the import again shidna be ae proablem.

Gif ye speceefied tae incluid histerie information, than foreby ye shid see information aneat the eedits in the 'histerie' o the importit pages, n in the uiser contreebutions. The eedits will na kith in 'recent chynges' (na poseetioned at the time o the oreeginal eedit, or at the time o importin). The effect will be siclik til ae ful histerie merge n it micht be quite difficult tae nail doun later whit eedits were importit.

Thaur's aen optie "Incluid aw templates", that'll import the templates that ar oan aen importit page. Gif this is na uised than thay will be reidairtit onless ae template o the same name exeests oan the tairget wiki. Gif this is uised, than templates o the same name oan the tairget wiki will be wruttenower n the import will incluid na juist the templates oan the importit page bit onie templates uised wiin thae templates aes weel. It's possible tae import ae collection o pages bi creautin ae page fer byordinar that transcluids thaim, n importin that page, wi the optie oan. Houever, this shid be dun verra carefulie or it micht ootcome in faur mae bein importit than ettled efter - again, it will import aw templates oan the transcluidit pages n aw templates enclaised wiin thae templates.

Uissfu appleecations o importin incluid:

  • whan ae page is muived til anither wiki n than eeditit thaur, hae the histerie thegather in the tairget wiki; this is pairticularlie uissfu gif the soorce page becomes mair difficult tae fynd cause o page muivs etc.
  • whan ae page is muived til anither wiki n delytit oan the soorce wiki, preserve the histerie.
  • in order tae hae templates that exeest oan anither wiki/subdomain.
  • oan certain waurks lik wikisoorce n wiktionar, tae muiv monieleidic content atween leid subdomains.

The carriein oot

Tae figure oot gif yer wiki haes the transwiki richt set up, n, whaur it's setup, til whit wikis, ye can speir yer wiki uisin the api.

The setup is communicatit in <param name="interwikisource" description="For interwiki imports: wiki to import from"> …

Or ye can owerluik the global setups at http://noc.wikimedia.org/conf/highlight.php?file=InitialiseSettings.php n luik in the section wgImportSources.

Gif thaur's nae transwiki setup, than seek yer local communitie's consensus tae hae the transwiki setup, n tae pinpoint whit wikis that ye micht want tae import fae. Speirins fer setup chynges shid be haunit in at Bugzilla: whau ye'd creaut ae new bug unner the Wikimedia: section. Ye'd be expectit tae airt til yer communitie's collogue in yer bug speirin.

Aen inspection o the API at yer local wiki wid displey the uisergroop richts aes weel, luik n see fer the <add> tag

Allocation o transwiki import richts

Maist transwiki richts ar allocatit til uisers follaein ae successfu collogue in thair communitie's wiki, n this is follaed bi ae speirin til stewards at Steward requests/Permission. Some wikis hae speired n been grauntit that local bureaucrats be able tae allou this richt, check wi yer wiki fer the seetuation there the nou.

Uplaid import

Hou tae export, n the format o exportit pages, is descreebit at Help:export. Normallie onie uiser can export wiki pages til ae file, but tae import pages intil ae wiki fae ae file, ye maun hae 'Seesop' (seestem operater) preevileges oan that wiki. Sae gif ye hae yer ain MediaWiki installation, than ye shid be able tae see the 'Special:Import' page thaur. Wiin the Wikimedia Foundation faimilie o wiki waurks, yinlie uisers wi the importupload uiser-richt can import pages intil ae wiki fae ae file; this yinlie incluids memmers o the "importer" groop n stewairds.

Tae import wiki pages fae yer computer, juist clap brouse tae fynd the file oan yer local file system.

Eeditin the import file

In the case o uplaid import, cause o the simple readable file format the XML file can easilie be eeditit atween exportin n importin. This shid be dun cannilie n wi honestie, ye can mak backdatit eedits n uise faus uiser names, n in combination wi delytion, ye can "chynge histerie". Uiss's o this eeditin incluid:

  • eikin ae note til the eedit ootline aneat the importin
  • chyngin uiser names n/or page names tae avoid name confleects (juist atween the title tags n atween the uisername tags or also in airtins n signatures)
  • chyngin namespace names intil the generic or the appleecable names (ditto)

Mynd that gif twa versions o the page hae the same timestamp (cause yin wis uplaidit wi the same timestamp aes ae preexeesting version), the later (imported) version will kith in the eedit histerie but na in the airticle itsel.

See mw:Manual:XML Import file manipulation in CSharp fer aen example o wairkin wi thir XML files in Visual Studio .NET C#.

Mergin histeries n ither compleecations

See mw:Manual:Importing_XML_dumps#Troubleshooting. aes weel

Gif the import incluids histerie information, n the eedits insnorlt ae uiser name that in the importin waurk is uised bi anither bodie, than uplaid import shid be uised, n the happenins o the uiser name in the XML file shid first be resteidit bi anither name, tae avoid doot. Gif the uiser name wis na uised yet in the importin waurk than the uiser contreebutions ar available oniehou, altho aen accoont is na autæmaticlie creautit.

Juist lik whan ae page is pointit til in aen airtin, n/or pit in ae URL, generic namespace names ar autæmaticlie chynged, n gif ae prefix is na ae namespace name the page will arrive in the main namespace. Houiver, e.g. "Meta:" micht be ignored (drappit) oan ae waurk that uises that prefix fer interwiki airtin. It micht be worth it tae chynge it in the XML file til "Project:" afore importin.

Gif ae page name awreadie exeests, than importin reveesions o ae page wi that name causes the page histeries tae be merged. Mynd that efter insertin ae reveesion atween twa exeestin reveesions in the page histerie, the chynge dun bi the uiser that did the neist eedit seems differant til whit it's actuallie been: tae see the actual chynge dun bi the uiser than ye hae tae tak the diff atween the twa awreadie exeestin reveesions, na the diff wi respect til the insertit reveesion. Syne this shidna be dun except tae reconstruct the true page histerie.

Ae reveesion is na importit gif ae reveesion o the same date, n exactlie the same time up til the seicont, exeests awreadie (beware that this disna seem tae happen in aw cases). In practice this happens yinlie whan the reveesion haes awready been importit afore, or whan the reveesion that ye were ettlin tae import wis importit the ither waa roond, or baith were importit fae ae third steid.

Aen eedit ootline micht speak o, n possiblie airt til, anither page. This micht be confusin whan the page haes been importit bit the tairget page haes na.

The eedit ootline disna autæmaticlie shaw that the page haes been importit, bit in the case o uplaid import that can be eikit til the eedit ootlines in the XML file afore importin. That can avoid some potential soorces o doot n/or confusion. Whan eeditin the XML file wi fynd/resteid, mynd that eikin ae tex til the eedit ootlines needs tae be distinct atween eedits that awreadie hae aen eedit ootline, hyne comment tags in the XML file, n thae wioot thir tags. Gif thaur's monie pairs o comment tags, yinlie the hintmaist is effecteeve.

Uiser contreebutions

Wioot proveesions fer uiser name conflicts, the uiser contreebutions leet shaws:

  • the eedits bi the bodie registerit unner the uiser name concerned oan the waurk
  • fer ilkin wiki that pages hae been importit fae, the eedits o importit pages afore import, bi the uiser that oan the soorce waurk haes the uiser name concerned

Gif at the time o import the page didna exeest oan the tairget steid yet, the twa can be disteenguisht bi comparin the time o import wi the time o the eedit.

Gif the uiser page n uiser tauk page dinna hae ae uiser contreebutions airtin in the page margin than the uiser is na registerit, sae aw o thair eedits ar importit.

Muckle-scale transfer

Fer ae muckle-scale transfer, somebodie wi suffeecient seestem preevileges can muiv data wiin the server, this is mair practical than sendin muckle XML files fae the server til ae uiser's local computer n than back til the server.

Muckle files can be rejectit fer twa raisons. The PHP uplaid leemit, foond in the PHP confeeguration file php.ini:

 ; Maximum allowed size for uploaded files.
 upload_max_filesize = 20M

N foreby the skaukt variable leemitin the size in the input form. Foond in the mediawiki soorce code, includes/specials/SpecialImport.php:

   <input type='hidden' name='MAX_FILE_SIZE' value='20000000' /> 

Perhaps ye shid chynge the follaeing fower directeeves in php.ini:

; Maximum size of POST data that PHP will accept.
post_max_size = 20M
max_execution_time = 1000  ; Maximum execution time of each script, in seconds
max_input_time = 2000	    ; Maximum amount of time each script may spend parsing request data
; Default timeout for socket based streams (seconds)
default_socket_timeout = 2000

Relatit pages

  • data dumps describes the maintenance screept maintenance/importDump.php that provides aen alternate import mechanism, bit haesna aye remained in wairkin order wi recent MediaWiki releases
  • Importer