(→‎Next templates to import: Image with comment templates)
Line 118: Line 118:
* [[Template:page number]]
* [[Template:page number]]
* [[Template:dubious]]
* [[Template:dubious]]
 
* [[Template:Who]]


+ask Teratornis.
+ask Teratornis.

Revision as of 23:50, 1 December 2012

Note, regexW skills needed for the xml editing.

  1. Export from source wiki, turning on "Include templates." Do not import this file.
  2. Determine the list of pages to be imported.
    1. Delete everything except the title lines and turn those titles into wikilinks. (One method: Remove all linebreaks - regex matching \r?\n should work on most systems - then regex replace .*?<title>([^<]+)</title> with \* [[$1]]\n and delete the last paragraph, which will not be a page title.)
    2. Paste into an Appropedia page (e.g. the Sandbox).
    3. Delete all the blue links from the list, until that only redlinks remain. Be 100% sure, to avoid overwrite problems. (This assumes that the blue links represent the same template working close enough to the same way on both wikis.)
    4. For each template in the list, add a /doc page below it. Replace (\* \[\[Template\:[^]]+)\]\] with $1]]\n$1/doc]]
    5. On second thoughts, this step may not be necessary as importing a non-existent page should simply fail with no action taken. Take this new list and paste it into Wikipedia's Sandbox. Click "Preview". Remove any redlinks. Preview to check. This is the list to be exported from the source wiki.
  3. Go to the export page on the source wiki
    1. Paste in the short list of red links (removing the square brackets)
    2. Turn off "Include templates"
    3. Turn on "Save as file"
    4. Export
  4. Edit the xml export file:
    1. Strip out interlanguage links, if any - step through each one to avoid false positives. This regex match code should get almost all, with few false positives: ^\[\[([a-z][a-z][a-z]?[\:\-]|simple\:)[^\]]+\]\](\r?\n)? and hit delete or replace with nothing.
    2. Change every username - for Wikipedia, assuming you've imported only the latest revision, change usernames to "(various editors @ en.wikipedia.org)". Regex: Match \<username\>[^\<]+\</username\> and replace with <username>(various editors @ en.wikipedia.org)</username>
  5. Double check there are no existing templates (e.g. do a quick search and eyeball each <title> line, comparing it to your memory - what you expected to be there.)
  6. Import here.
  7. Check the templates.
  8. Add attribution, e.g. by pasting in {{attrib wikipedia|{{FULLPAGENAME}} }} at the bottom of the "noinclude" section. (This code assumes the template is from Wikipedia and the name is unchanged.)


Issues

NEXT STEPS:

Late 2012:

  • Add {{attrib wikipedia|{{subst:PAGENAME}}}} to each imported template. (On template page before last noinclude, for greatest reliability... Or on the /doc page so it's on every relevant page, including the /doc page when viewed directly. A bit fiddly to deal with the cases of no /doc, but there are very few such cases.)
  • Template:SometimesW looks wrong

Weird addition of content (early revisions of mainspace content) to failed templates, e.g. STAR-TIDES page at Template:Sometimes/doc.

Next templates to import

Do next mass import after server upgrade to avoid charset errors, and preferably when we have enabled either multiple image upload, or displaying images from Commons.

+ask Teratornis.

Categories/groups of templates (import category pages, but also need to add each template, I think):

Import then rename:

Update, & edit doc page after import:

Tools or processes to import and/or adapt, with their templates. (This will take work to identify & list the templates.):

Manual import & edit:

  • Think about adapting wikipedia:Template:WikiProject Norway (or any such talk page notice) - worth trying one or two such groups on Appropedia? As long as it's not relying just on conventional talk pages for communication - need some combination of a bot to update talk pages (opt-in), alerts for changes to that page (separate from standard watchlist alerts), an active facilitator, a related academic program working on the area, and/or a social media person promoting the area and the collaboration.
Cookies help us deliver our services. By using our services, you agree to our use of cookies.