Wikipedia:Miscellany for deletion/Mass-created portals based on a single navbox/Selection process

Selection process for this nomination
I offered to document the process by which I selected a list of articles for this nomination. @Pldx1 said that would be useful, so here it is.

Note that it requires a programmer's text editor and WP:AWB. Some modest proficiency is needed with AWB's list-making features, but no great wizardry.

I took 9 steps. Several of the steps could be merged, but I prefer to do tasks like this one step a time.


 * 1) View contribs. Display the user contributions by @The Transhumanist, selecting portal namespace, Only show edits that are page creations, 500 pages at a time — start here Note that this will no longer show the same results as when I did it, because some pages have since been deleted. However, I think that if you use the same cutoff dates, the end results should be identical, because the pages which have since been deleted were excluded in later steps.  The last entry in the 4 pages I screenscraped was Portal:Bowerbirds, created on 22:41, 12 September 2018‎.
 * 2) Screenscrape. Copy the first 4 screenfulls to a text editor (I was using Notepad++).
 * 3) Strip irrelevancies.  On each line strip off everything except the page name. This can be done accurately thanks to hidden characters which were included in the screengrabs. I can't display the hidden chars (value U+200E: LEFT-TO-RIGHT MARK [LRM]) so I have replaced them with X, but with that modification the two regexes are:   and
 * 4) Linkify. Use a regex to convert each line to a link:  .  Save list as a file
 * 5) Remove duplicates.  Load list into WP:AWB. Use List->Remove duplicates.  Save list
 * 6) Remove redirects. Use WP:AWB in list-making ("pre-parse") mode to skip any page which matches the regex  . Save list.
 * 7) Remove pages already tagged for MFD. Use AWB in list-making ("pre-parse") mode to skip any page which matches the regex  . Save list.
 * 8) Remove non-automated. Use AWB's "List comparer" tool" to keep only pages which transclude Template:Transclude list item excerpts as random slideshow.  Save list.
 * 9) Keep only single-navbox selections. I used an AWB custom module (see my module to identify pages where the only unnamed parameter for Transclude list item excerpts as random slideshow is a single template; other pages were skipped.  Save list.

That final list after using the custom module was the one I included with the nomination, after a dozen random spot-checks. The module is at Miscellany for deletion/Mass-created portals based on a single navbox/AWB module.

When I initially made the first list, I stupidly omitted step #8 "Remove non-automated", thinking that the module would handle it, but I had somehow removed the relevant "Skip" command from the module. When @Northamerica1000 kindly notified me of a false positive in the list I saw the error, and ran that step, which eliminated 36 pages from the initial set of 1,426.

I would welcome any checks by others on this process. After fixing my initial glitch which led to the 36 false positives, I have re-run it several times with the same results. It would be nice to know that others had assessed my techniques and code ... and either replicated the result or identified errors to be fixed. -- Brown HairedGirl (talk) • (contribs) 20:11, 8 April 2019 (UTC)


 * Thanks for this listing. I will do my own home work and report here. Pldx1 (talk) 20:42, 8 April 2019 (UTC)

Tracking categories enabled
The automated portals, including those nominated for deletion here, use Module:Excerpt slideshow to build their "Selected general articles" list. I have modified that module so that portals which build their article lists from only one two two templates will be categorised in these two tracking categories:
 * Category:Automated portals with article list built solely from one template (population: )
 * Category:Automated portals with article list built solely from two templates (population: )
 * Usages with more than two templates are not currently tracked.
 * Note that the tracking was enabled 9 April 2019‎ at 15:28 UTC, and it will probably take a few days for all the portal pages to update. -- Brown HairedGirl (talk) • (contribs) 15:51, 9 April 2019 (UTC)
 * Update: as far as I can tell, all pages have been refreshed, and the figures are up to date. -- Brown HairedGirl (talk) • (contribs) 18:08, 9 April 2019 (UTC)
 * Update: as far as I can tell, all pages have been refreshed, and the figures are up to date. -- Brown HairedGirl (talk) • (contribs) 18:08, 9 April 2019 (UTC)