Difference between revisions of "Issue:Subpages as sublists"

Latest revision as of 12:51, 22 April 2009

Problem

I have a statement like this

   <dpl>
   category=Mycenean
   </dpl>

Now some of the pages in the category are subpages so I get a display like this

    * [[Greek/Mycenean]]
    * [[Greek/Mycenean/Texts]]
    * [[Greek/Mycenean/Writing]]
    * [[Greek/Mycenean/Writing/Syllabary]]

• Greek/Mycenean
• Greek/Mycenean/Texts
• Greek/Mycenean/Writing
• Greek/Mycenean/Writing/Syllabary

I would like them to appear as an hierarchical listing

    * [[Greek]]
    ** [[Greek/Mycenean|Mycenean]]
    *** [[Greek/Mycenean/Texts|Texts]]
    *** [[Greek/Mycenean/Writing|Writing]]
    **** [[Greek/Mycenean/Writing/Syllabary|Syllabary]]

• Greek
  • Mycenean
    • Texts
    • Writing
      • Syllabary

N.B. That [[Category:Greek]] would have further subpages [[Greek/Writing]], [[Greek/Dialects]] and whatever which shouldn't appear in a listing of category=Mycenean, but [[Greek]] itself should, as the top page in the category=Mycenean hierarchy is itself a subpage of that page.

As for constructing code which translates the current wikitext into the desired one this Perl regex does the trick, except that they don't add the [[Greek]] top node (There would need to be some check to see if the category named in a category= parameter is itself a subcategory and if so add the supercategory/-ies):

s{^\*[ ](\[\[.+?([^/]+)\]\])}{ 
    $l = $1; 
    $c = ($l =~ s#/#/#g); ## Count the number of slashes in the link and add 1.
    $l =~ s#([^/]+)\]\]$#$1\|$1\]\]#;  ## Copy link text
    "\*" x $c  ." ".$l;  ## Print out the right number of asterisks and the link
}egm;

This is the concept. A bit harder of course if done on HTML rather than wikitext.

Melroch 10:51, 21 April 2009 (UTC)

Reply

May be someone (you??) could write a separate little extension which takes the wiki text of a flat enumeration list and transforms it into an indented enumeration list; the symbol which is being used to recognize the hierarchy (in your case the '/') could be an optional parameter. If there was such an extension it could easily be used from within DPL I think. Gero 06:51, 22 April 2009 (UTC)

I guess I could if I was as comfortable with PHP as I am with Perl. I'm a scholar who has learned to use some Perl to explore my data, not a programmer. There is also still the problem of superpages (which I mistyped as supercategories) which are not themselves included in the search match. I think it is not really so easy as my example regex to modify the wikitext makes it appear. To get a proper output one still has to gather the hierarchy into a datastructure (I'm thinking a hash of hashes) from which the new HTML list can be constructed, so that one doesn't have to trust that he input is a well-formed, sorted list of pages which all belong to the same subpage hierarchy. I'll write a proof-of concept in Perl and then see if I can translate it into PHP or find someone to do it for me. Melroch 11:51, 22 April 2009 (UTC)