The feedback request service is asking for participation in this request for comment on Talk:Useful idiot. Legobot (talk) 04:29, 3 January 2018 (UTC)[reply]

Is that you DumbBOT?

—usernamekiran(talk) 19:59, 3 January 2018 (UTC)[reply]

import bz2
from lxml import etree
file_path = 'enwiki-latest-pages-articles.xml.bz2'

tag_count = 0

with bz2.open(file_path, 'rb') as file:
    context = etree.iterparse(file, events=('end',), tag='{http://www.mediawiki.org/xml/export-0.10/}page')
    if tag_count % 1000:
        print(tag_count)
    for _, elem in context:
        tag_count += 1

        elem.clear()

        while elem.getprevious() is not None:
            del elem.getparent()[0]

print("Total page tags:", tag_count)

@Qwerfjkl: Hi. yes, you are right. The pod gets OOMKilled. It was confusing, not much information was around, and it was pain in the butt. When I was having this (similar) issue, I solved it by loading the xml file in parts, and then breaking the file in chunks based on the closing tags. Only after breaking in chunks, I could process the dump. As soon as I get to the computer, I will upload all the relevant programs to github repository, and will let you know. But that may take 12ish hours from now. —usernamekiran (talk) 18:25, 2 November 2023 (UTC)[reply]

It seems to be split into 27 chunks already (e.g. enwiki-20231101-pages-articles1.xml-p1p41242.bz2), I might try seeing of those work. — Qwerfjkltalk 19:14, 2 November 2023 (UTC)[reply]

@Qwerfjkl: Hi. Sorry, I forgot to mention, we need to unzip the bzip file before processing it to conserve the memory/RAM. The unzip/extraction process takes a lot time though, and the uncompressed file is very big in size, so - on toolforge - it is recommended to delete it after being done with it. But there seems to be something wrong/missing with the way I broke the unzipped xml file into chunks. The results of my next script were not very much accurate. Kindly let me know if there was something wrong with it. Also, please dont mind my rudimentary skills of python. This script also could use better error handling. github repository. —usernamekiran (talk) 12:35, 3 November 2023 (UTC)[reply]

If you look at my code above, I used the streaming decompression library bz2 and lxml for the parsing. So I don't think decompression is necessary. From what I can tell it seems to run okay. — Qwerfjkltalk 14:47, 3 November 2023 (UTC)[reply]

• Maybe following script would work, I couldn't test it, as I am on a different computer.

import bz2
from lxml import etree

file_path = 'enwiki-latest-pages-articles.xml.bz2'
tag_count = 0

with bz2.open(file_path, 'rb') as file:
    context = etree.iterparse(file, events=('end',), tag='{http://www.mediawiki.org/xml/export-0.10/}page')

    for _, elem in context:
        tag_count += 1
        elem.clear()

        while elem.getprevious() is not None:
            del elem.getparent()[0]

        if tag_count % 1000 == 0:
            print("Processed", tag_count, "pages")

print("Total page tags:", tag_count)

@Qwerfjkl: —usernamekiran (talk) 15:12, 3 November 2023 (UTC)[reply]

Correct me if I'm wrong, but that seems to essentially be the same as my script. — Qwerfjkltalk 16:29, 3 November 2023 (UTC)[reply]

@Qwerfjkl: yeah you are right. When I first looked at your script, I thought (mistakenly) that something was wrong wrong with the counting method. So I pasted your script in windows notepad, and then got busy in IRL, when I came back I started to edit your script. In short, there is no issue with your script. Sorry about the muck-up. Did you try breaking the original/unzipped xml file into smaller xml files, and then parsing these files? "etree.iterparse" wouldnt consume much memory, and the script is also removing the processed element(s) from memory, I am not sure why this might be failing. are you sure it is OOMKilled? what do the .err/.out files say? Maybe lxml dependency was not installed on your toolforge correctly? —usernamekiran (talk) 16:53, 3 November 2023 (UTC)[reply]

When running the script it literally just errors "Killed." Nothing in the script actually runs afaict, or at least, none of the print statements are executed and outputed. — Qwerfjkltalk 17:10, 3 November 2023 (UTC)[reply]

• @Qwerfjkl: Hi. Currently I'm out of town, with no computer. I will look into this as soon as I get back, hopefully we can find the issue. —usernamekiran (talk) 13:53, 5 November 2023 (UTC)[reply]

  @Qwerfjkl: Hi. I am back home. Did you get any new information regarding the parsing? —usernamekiran (talk) 18:25, 12 November 2023 (UTC)[reply]

  No, I've been taking a break to work on some personal coding projects. — Qwerfjkltalk 21:20, 12 November 2023 (UTC)[reply]

  @Qwerfjkl: Hi. I tested your script through cron. The only change I made was to save the progress to a text file instead of printing it. Worked fine for me. —usernamekiran (talk) 19:07, 15 November 2023 (UTC)[reply]

  I restarted the task, the start, and end will be documented at https://test.wikipedia.org/wiki/User:KiranBOT/MOSTREFS/log —usernamekiran (talk) 19:21, 15 November 2023 (UTC)[reply]

  Thanks, saving the output to a text file seems to have worked for me. — Qwerfjkltalk 20:36, 15 November 2023 (UTC)[reply]

  I've being indexing the pages with the code below, but it seems to run out of memory after about 3 million pages (at least, it's killed; I assume memory is the problem). Do you know any way to fix this?
  

  def init_memory_index_of_articles(file_path):
      articles = set()
      redirects = dict()
      manually_excluded_pages = set()
  
      printm("Building index of Articles ...")
  
      with bz2.open(file_path, 'rb') as file:
          context = etree.iterparse(file, events=('end',), tag='{http://www.mediawiki.org/xml/export-0.10/}page')
          count = 0
          for _, elem in context:
              count +=1
              if count % 100000 == 0:
                  printm(count)
              title_element = elem.find('{http://www.mediawiki.org/xml/export-0.10/}title')
              text_element = elem.find('{http://www.mediawiki.org/xml/export-0.10/}revision/{http://www.mediawiki.org/xml/export-0.10/}text')
              
              # Ensure the title and text elements are found
              if title_element is not None and text_element is not None:
                  title = title_element text
                  text = text_element text
                  if not text:
                      continue
  
                  # Check pagetype:
                  redirect_element = elem.find('{http://www.mediawiki.org/xml/export-0.10/}redirect')
                  if redirect_element is not None:
  
                      redirect_target = redirect_element.get('title')
                      redirects[title] = redirect_target
  
                      # printm(f'Processed {title}, is redirect')
                  elif '((disambig' in text.lower():
                      manually_excluded_pages.add(title) # exclude dabs
                      # printm(f'Processed {title}, is dab')
                  else:
                      articles.add(title)
                      # printm(f'Processed {title}, is article')
              
              # Clear the element from memory
              elem.clear()
              # Also eliminate now-empty references from the root node to elem
              while elem.getprevious() is not None:
                  del elem.getparent()[0]
          
      printm("Completed creating memory index ...")
      return redirects, articles
  

  — Qwerfjkltalk 12:55, 17 November 2023 (UTC)[reply]

  @Qwerfjkl: Hi. Currently I do not have dump in my toolforge userspace, I wanted to copy it from your userspace, I cant find it in your userspace either. Maybe it is not public, or are you trying to process the file remotely? Generally speaking, manual script running from terminal (from virtual env) or from bastion/shell for resource heavy, or time consuming tasks is not recommended. scheduled jobs from yaml file is more efficient (than one-off jobs). If you create a yaml file as described here, you would get better/automated logging. For one-off, or testing purposes, I use crons similar to "55 19 18 11 *". I am saying this because I am getting an inkling that you are running the script from terminal. Maybe thats why it is accumulating memory even though you are clearing each element after processing it. Other than that, I cant think of anything else for now. Kindly let me know if you can think of anything. —usernamekiran (talk) 16:35, 17 November 2023 (UTC)[reply]

  I noticed it was stored on the Toolforge servers so I switched to using
  

  file_path = '/mnt/nfs/dumps-clouddumps1002.wikimedia.org/enwiki/latest/enwiki-latest-pages-articles.xml.bz2'
  

  I have been running from terminal, yes. I'll try switching to yaml files and see if that helps. — Qwerfjkltalk 17:15, 17 November 2023 (UTC)[reply]

  It turns out the very large dictionary was causing the memory issue. I'm going to try a different approach, using a multistream version. — Qwerfjkltalk 14:22, 18 November 2023 (UTC)[reply]

  @Qwerfjkl: Hi. I had worked on the dump script almost an year ago, I think I had tried the multistream approach too, but I changed my approach halfway. To circumvent the memory issues, you should save the output to multiple files eg redirects_1.json, redirects_2.json articles_1.json and so on. I think you should keep the maximum number of entries to these files either 5000 or 10000, definitely not more than 10k. —usernamekiran (talk) 10:48, 21 November 2023 (UTC)[reply]

┌───────────────────────────┘
So you suggest creating 2300 json files? — Qwerfjkltalk 15:06, 21 November 2023 (UTC)[reply]

@Qwerfjkl: yes. If you want/need, then at the end of the script, or through another script you can combine all these files together, and delete the ~2400 files. —usernamekiran (talk) 16:23, 21 November 2023 (UTC)[reply]

But surely reading those files for usage would just overload the memory, especially if combined? — Qwerfjkltalk 16:36, 21 November 2023 (UTC)[reply]

@Qwerfjkl: not much, by my estimation, the complete operation of merging the files would need around 550 MiB. —usernamekiran (talk) 20:40, 21 November 2023 (UTC)[reply]

I see, I'll try that. My question is, I suppose, if we can load it by reading json files, why can't we store it by reading xml dumps? — Qwerfjkltalk 21:24, 21 November 2023 (UTC)[reply]

@Qwerfjkl: it is related to memory accumulation, and python's garbage collector. and we will be using streaming approach to merge the files, which should be memory efficient. I cant be 100% sure though. I have started a test run on toolforge a few minutes ago, lets see how it goes (so far, the program is using around 40Mi). As I was not sure why do you need the data, and I had to make some minor changes in the script, the output would be in txt files, in following format:

'Anarchism',
'Albedo',
'A',

and

'AccessibleComputing': 'Computer accessibility',
'AfghanistanHistory': 'History of Afghanistan',
'AfghanistanGeography': 'Geography of Afghanistan',

I hope that format is okay for you. —usernamekiran (talk) 12:52, 22 November 2023 (UTC)[reply]

All I need is, for a given page title, to check whether it exists, and if it does, is it a redirect (and if so what is the redirect target). — Qwerfjkltalk 15:29, 22 November 2023 (UTC)[reply]

@Qwerfjkl: the script finished successfully with created the file for articles with 378 MBs, and files for redirects for 555 MBs. and the highest memory value consumed by the script was 108.4 MiB. —usernamekiran (talk) 18:54, 22 November 2023 (UTC)[reply]

My issue would be when reading the files, especially if they were combined (for my purposes, I think I'll just represent redirects as 'RedirectTitle': 'RedirectTarget' and articles as 'ArticleTitle: '', in the same file). — Qwerfjkltalk 19:36, 22 November 2023 (UTC)[reply]

Okay, I've implemeted this. I have a 2 GB JSON file containing the relevant data, but when I try to load it into memory using JSON.load(file) after about a minute it gets killed. — Qwerfjkltalk 15:52, 29 November 2023 (UTC)[reply]

I'm currently trying storing the data on a SQL database. — Qwerfjkltalk 17:39, 9 December 2023 (UTC)[reply]

@Qwerfjkl: Hi, sorry for the delayed reply, I was feeling a little under the weather. Unfortunately, I do not have any experience with databases, so I cant help you much. What are you trying to achieve, I mean whats your goal/target? Maybe I/we can come-up with some workaround bypassing the use of SQL? —usernamekiran (talk) 10:32, 16 December 2023 (UTC)[reply]

I'm trying to restart the Missing Redir4ects Project. You can see this for an example of the output I'm trying to get. — Qwerfjkltalk 11:31, 16 December 2023 (UTC)[reply]

In case you're interested, I made a website for this at toolforge:missingredirectsproject. — Qwerfjkltalk 18:21, 17 January 2024 (UTC)[reply]

@Qwerfjkl: Hi, thanks for the update, it is appreciated a lot. May I ask how did you manage the previous issues? —usernamekiran (talk) 02:52, 19 January 2024 (UTC)[reply]

Like I said, using a sql database. — Qwerfjkltalk 07:04, 19 January 2024 (UTC)[reply]

	Greetings, Usernamekiran. You are receiving this notification because you've agreed to consider endorsing prospective admin candidates identified by the process outlined at Administrators without tools. Recently, the following editor(s) received this distinction and the associated endearing title: Clovermoss
Thank you for supporting this effort. Your contributions are an integral part of overall success, and an example for others to follow. To stop receiving these notifications, remove your name from the list.

	Greetings, Usernamekiran. You are receiving this notification because you've agreed to consider endorsing prospective admin candidates identified by the process outlined at Administrators without tools. Recently, the following editor(s) received this distinction and the associated endearing title: MPGuy2824
Thank you for supporting this effort. Your contributions are an integral part of overall success, and an example for others to follow. To stop receiving these notifications, remove your name from the list.

New year, new scripts. Welcome to the 23rd issue of the Wikipedia Scripts++ Newsletter, covering around 39% of our favorite new and updated user scripts since 24 December 2021. That’s right, we haven’t published in two years! Can you believe it? Did you miss us?

Got anything good? Tell us about your new, improved, old, or messed-up script here!

Featured script

RefRenamer by Nardog is this issue's featured script. For whatever reason, the foundation will only give us better reference names in VisualEditor if we wish for it in the next survey. Nardog to the rescue! This script has a ton of options to satisfy almost anyone who wishes to use better refnames than ":0" and ":418".

Updated scripts

• Nardog: MoveHistory can now look up the move history of any page from anywhere on the site.
• User:Enterprisey/diff-permalink now has a version that copies the Special:Diff link with square brackets.
• User:GhostInTheMachine/WatchlistTidy now moves the Active Filters panel to the right so that the Show/Hide button doesn't constantly jump.
• User:Alexander Davronov/HistoryHelper has now become stable with some bugfixes and features such as automatically highlighting potentially uncivil edit summaries and automatically pinging all the users selected.
• Wikipedia:AutoEd/unicodehex.js will no longer replace legacy MediaWiki dot-entities (dot followed by two uppercase alphanumeric characters) in article links, which caused several links to become broken.

Improve a script

• Many of our favorite scripts such as User:Fred Gandt/userResourceManager, User:Guarapiranga/search-shortcuts and User:Bradv/Scripts/Superlinks haven't been updated for Vector 2022. You could be the one to change that!
• User:Elominius/gadget/media timer works, but its UI looks alien. Someone could improve it...
• To a lesser extent, the same goes for User:PrimeHunter/Search sort.js. I wish someone would integrate the sorts into the sort menu instead of adding 11 portlet links.
• We want...

• ...a mark-as-read button for Watchlyst Greybar Unsin
• ...a tool to automatically fix CAT:BROKENREFs, preferably a fork/addition to User:TheJJJunk/ARA
• ...a fix/fork of User:DannyS712/SectionRemover.js to make it work

Requested scripts

• A WP:3O requesting and responding tool, similar to Terasil's classic edit request tool
• A "bytes removed"/"added" filter on the user contribs page
• A review tool for WP:proposed deletions
• Mass patrol new pages by a user for use in WP:RFP/A
• ...and many more, all available at Wikipedia:User scripts/Requests

New scripts

• Aaron Liu: Watchlyst Greybar Unsin is a rewrite of Ais's Watchlist Notifier with modern APIs and several new features such as not displaying watchlist items marked as seen (hence the name), not bolding diffs of unseen watchlist elements which doesn’t work properly anyways, displaying the rendered edit summary, proper display of log and creation actions and more links.
• Alexis Jazz: Factotum is a spiritual successor to reply-link with a host of extra features like section adding, link rewriting, regular expressions and more.
• User:Aveaoz/AutoMobileRedirect: This script will automatically redirect MobileFrontend (en.m.wikipedia) to normal Wikipedia. Unlike existing scripts, this one will actually check if your browser is mobile or not through its secret agent string, so you can stay logged in on mobile! Hooray screen estate!
• Chlod:
  • Deputy is a first-of-its-kind copyright cleanup toolkit. It overrides the interface for Wikipedia:Contributor copyright investigations for easy case processing. It also includes the functionality of the following (also new) scripts:
    • Attribution Notice Template Editor (ANTE) graphically modifies content attribution notices. It is the successor to User:Chlod/Scripts/CopiedTemplateEditor and supports more templates than ((copied)) ones.
    • Infringement Assistant allows a user to hide sections of a page for suspected copyright violations and report the page to the appropriate Wikipedia:Copyright problems noticeboard.
• User:Elominius/gadget/diff arrow keys allows navigation between diffs with the arrow keys. It also has a version that requires holding Ctrl with the arrow key.
• Enterprisey
  • diff-permalink-2 allows quickly copying wikilinks to diff and permalink pages from those pages, a variation of existing script User:Enterprisey/diff-permalink
  • archiver is a fork of Σ's Archiver with several bugfixes.
  • /parent-cats highlights categories that are a subcategory of a category already included in a page's categories.
• User:Guarapiranga/accessKeysCheatSheet adds an access key shortcut to display all shortcuts based on code provided by Nardog and NguoiDungKhongDinhDanh.
• User:HungKhanh0106/ProperDisplay loads a bunch of fonts to add correct rendering supports for "esoteric" languages.
• User:Jeeputer/highlightPiped will make piped links irresistible (or else).
• User:Jerome Frank Disciple/curlyfixer.js adds an WP:AutoEd function that straightens curly quotes per MOS. That's right! Such a module was only made in the 14th year of AutoEd's existence! Configure away!
• Frequently link to Wikipedia on your websites yet find generating CC-BY credits to be such a hassle? Say no more! User:Luke10.27/attribute will automatically do it for ya and copy the credit to yer clipboard.
• User:MPGuy2824/MoveToDraft, a spiritual successor (i.e. fork) to Evad37's script, with a few bugs solved, and a host of extra features like check-boxes for choosing draftification reasons, multi-contributor notification, and appropriate warnings based on last edit time.
• Nardog
  • /CopyCodeBlock: one of the most important operations for any scripter and script-user is to copy and paste. This script adds a copy button in the top right of every code block (not to be confused with <code>) that will, well, copy it to your clipboard!
  • /RCMuter: Hide specific users, perhaps User:MediaWiki message delivery, from the Watchlist and Recent Changes.
  • /VitalTopicon makes you see the vital article topicon on WP:vital articles.
• m:User:NguoiDungKhongDinhDanh/AceForLuaDebugConsole.js adds the Ace editor (a.k.a. the editor you see when editing JS, CSS and Lua on Wikimedia wikis) to the Lua debug console. "In my opinion, whoever designed it to be a plain <textarea> needs to seriously reconsider their decision."
• Novem Linguae
  • DraftCleaner fixes formatting of new articles.
  • GANReviewTool quickly and easily closes good article nominations.
  • ReviewStatus displays whether or not a mainspace page is marked as reviewed.
  • SpeciesHelper tries to add the correct speciesbox, category, taxonbar, and stub template to species articles.
• User:Opencooper/svgReplace and Tol's fork replaces all rasterized SVGs with their original SVG codes for your loading pleasures. Tell us which one is better!
• User:Phlsph7/HighlightUnreferencedPassages adds a button to highlight all unreferenced paragraphs in reds, and can even do it automatically.
• Still hate Vector 2022's floating TOC? User:Phlsph7/UnfoldedNumberedTOC(Vector2022) might make you give it a try.
• Quite a bit of pages at Special:WantedCategories either exist now or got removed. User:Qwerfjkl/scripts/hideUnwantedCategories automatically hides those.
• User:SuperHamster's Wikipedia:Tools/View it! generates a gallery of images for a given subject on any Wikimedia project utilizing Structured Data on Commons. Images can then be added as needed.
• Terasail
  • ArticleInfo displays page information at the top of the page, directly below the title.
  • /HeaderIcons takes away the Vector 2022 user dropdown and replaces it with all of the icons within, top level, right next to the Watchlist. One less click away! There's also an alternate version that uses text links instead of icons.
• ...and any more, all available at WP:US/L.

Due to budget and procrastination limitations, we've only had time to review the first ~39% of the scripts since last issue. Help us out here! Aaron Liu (talk) 13:24, 29 December 2023 (UTC)[reply]

---

• About the newsletter
• Archive
• Subscribe or unsubscribe
• Discuss this issue

	Greetings, Usernamekiran. You are receiving this notification because you've agreed to consider endorsing prospective admin candidates identified by the process outlined at Administrators without tools. Recently, the following editor(s) received this distinction and the associated endearing title: Sdkb
Thank you for supporting this effort. Your contributions are an integral part of overall success, and an example for others to follow. To stop receiving these notifications, remove your name from the list.

Hello. I've blocked your bot as it is doing weird things - at Wikipedia:In the news/Posted/December 2004 the text gets smaller and smaller (and contained a ((delete)) which is how I noticed) but more concerningly at Wikipedia:In the news/Posted/January 2005 it's added a NSFW image. I presume that's vandalism it's importing from elsewhere, but regardless, it doesn't strike me as something a bot should be doing. Happy for anyone to unblock so long as you know what it is doing. SmartSE (talk) 19:09, 30 December 2023 (UTC)[reply]

@Smartse: Hello. I'm currently on mobile, so I'm not sure about the text getting smaller. The vandalism was expected, and I was going to remove it, and clean-up the other formatting issues after 12ish hours from now. In short, the bot is archiving entries of additions, updates, and removals from Template:In the news. Would you kindly unblock the bot? I will repair all the pages in 12 to 14 hours from now. These archive pages are recently created by the bot, and not watched by anyone. —usernamekiran (talk) 19:30, 30 December 2023 (UTC)[reply]

So you knew it was going to do this? That doesn't seem like a great idea. Why aren't you either filtering it as you go, or more sensibly, dumping the output locally and then cleaning it up before posting it. There doesn't seem to be any need for the bot to doing the editing. I'm not going to unblock it if it's going be carry on doing that and you're not around to fix it. SmartSE (talk) 19:45, 30 December 2023 (UTC)[reply]

@Smartse: Not exactly, I was expecting garden variety vandalism, not this type of vandalism. I terminated the program responsible for the archiving task. There is another task ongoing, that's why I'm requesting for the unblock. Also, I had tried the approach for working on the pages locally like you suggested, I will perfect that method, and upload the tidied up pages. —usernamekiran (talk) 19:57, 30 December 2023 (UTC)[reply]

Ok thanks. I will unblock. SmartSE (talk) 21:24, 30 December 2023 (UTC)[reply]

thanks —usernamekiran (talk) 22:08, 30 December 2023 (UTC)[reply]

I came here to post about these syntax errors. I am glad to hear that you are planning to fix them. Most of the newly created pages have errors; some of them have many dozens of them. You can see the error list on each page's Page information page, under "Lint errors". I cleaned up a couple of pages (see this and this), but the bot continues to edit and create pages, so I assume that you have a plan. Feel free to ping me if you have trouble fixing any of the errors. – Jonesey95 (talk) 23:38, 30 December 2023 (UTC)[reply]

discussion ongoing. —usernamekiran (talk) 00:26, 31 December 2023 (UTC)[reply]

Hi. How do I edit and is there stuff or articles about scp --NobleChernobyl (talk) 19:53, 31 December 2023 (UTC)[reply]

Happy New Year!

Usernamekiran,
Have a prosperous, productive and enjoyable New Year, and thanks for your contributions to Wikipedia.
Chris Troutman (talk) 20:04, 31 December 2023 (UTC)[reply]

   Send New Year cheer by adding ((subst:Happy New Year fireworks)) to user talk pages.

thank you very much Chris, it is appreciated a lot. I hope the same for you, and your family. —usernamekiran (talk) 09:21, 1 January 2024 (UTC)[reply]

News and updates for administrators from the past month (December 2023).

Administrator changes

Clovermoss

Dennis Brown

• Julia W
• Marasmusine
• PBS

CheckUser changes

• Aoidh
• HJ Mitchell
• Sdrqaz
• Spicy
• ToBeFree
• Vanamonde93
• Z1720

Maxim

• Enterprisey
• Izno
• SilkTork

Oversighter changes

• Aoidh
• Firefly
• Sdrqaz
• ToBeFree
• Z1720

Maxim

• Enterprisey
• Izno
• SilkTork

Arbitration

• Following the 2023 Arbitration Committee elections, the following editors have been appointed to the Arbitration Committee: Aoidh, Cabayi, Firefly, HJ Mitchell, Maxim, Sdrqaz, ToBeFree, Z1720.
• Following a motion, the Arbitration Committee rescinded the restrictions on the page name move discussions for the two Ireland pages that were enacted in June 2009.
• The arbitration case Industrial agriculture has been closed.

Miscellaneous

• The New Pages Patrol backlog drive is happening in January 2024 to reduce the backlog of articles in the new pages feed. Currently, there is a backlog of over 13,000 unreviewed articles awaiting review. Sign up here to participate!