Inductiveload User Area
Main User Page Talk Page Talk archives Contributions

WELCOME to my user page. Have a look around my galleries and contributions if you like, or leave messages on my talk page. It is also worth checking in the IRC channels if you need me quickly. I am usually present in the main Wikisource channel, #wikisource. Come and have a chat!

If you can suggest improvements to my own work, tell me. Don't let any poor quality work hang around!

Wikisource user page Commons user page Wikibooks user page Wikipedia user page

Languages: English (native), French (basic), German (basic), Mandarin (basic), Mercan (like, totally fluent), Python (passable), C++ (basic), Lua (basic)


Go, go, go, said the editor: online kind
Cannot bear very much reality.

The Monthly Challenge for October contains 33 works.
This month:
  • Pages processed: 2377 (118.8% of target)
  • Avg. pages/day: 140
  • Yesterday: 84

Last month:
  • Pages processed: 2606 (130.3% of target)
  • Avg. pages/day: 87
Wikisource laurier.svg

Wikisource speaker icon.svg

Ask me to do thingsEdit

I will need information:

Tools and scriptsEdit

User preferences and custom javascripts:

Commons scriptsEdit

Popups ReloadedEdit

Popups, but way better.

mw.loader.load('//en.wikisource.org/w/index.php?title=User:Inductiveload/popups_reloaded.js&action=raw&ctype=text/javascript');
mw.loader.load('//en.wikisource.org/w/index.php?title=User:Inductiveload/popups_reloaded.css&action=raw&ctype=text/javascript', "text/css");

Quick AccessEdit

Keyboard-driven tool access

mw.loader.load('//en.wikisource.org/w/index.php?title=User:Inductiveload/quick_access.js&action=raw&ctype=text/javascript');

Cleanup (Alpha)Edit

This is a much more in-depth version of User:Samwilson/PageCleanUp.js that includes hundreds of regexes for scannos that have unambiguous or almost-certain corrections. For example, no work in English ends in -abte: this is almost certainly -able

This is an alpha-level tool. Configuration, especially, is likely to change.

mw.loader.load('//en.wikisource.org/w/index.php?title=User:Inductiveload/cleanup.js&action=raw&ctype=text/javascript');

This tool will work with default configurations, but it is more reliable with configurations.

Page CarouselEdit

Quick buttons to load the next/previous pages of a book.

mw.loader.load('//en.wikisource.org/w/index.php?title=User:Inductiveload/page carousel.js&action=raw&ctype=text/javascript');

Save/Load ActionsEdit

Run Javascript actions on page save or page load. Can be used to implement custom text transforms, for example from [[Link/Foo|Foo]] to [[Link/Foo|Foo]].

This tool probably needs configuration in your JS. Just adding it won't do anything use. Consult the documentation for more information.

Preview markupEdit

mw.loader.load('//en.wikisource.org/w/index.php?title=User:Inductiveload/show_markup.js&action=raw&ctype=text/javascript');
mw.loader.load('//en.wikisource.org/w/index.php?title=User:Inductiveload/show_markup.css&action=raw&ctype=text/javascript', "text/css");

Maintenance Wizard and ReplacerEdit

Perform maintenance without going into edit mode.

// maintain script has no purpose in special
if (mw.config.get("wgCanonicalNamespace") !== "Special") {
  mw.loader.using(['ext.gadget.utils-difference', 'mediawiki.util', 'mediawiki.api',
      'oojs-ui-core', 'oojs-ui-windows', 'oojs-ui-widgets']).done(function() {
    mw.loader.load("/w/index.php?title=User:Inductiveload/maintain.js&action=raw&ctype=text/javascript");
    mw.loader.load("/w/index.php?title=User:Inductiveload/maintain-ws-tools.js&action=raw&ctype=text/javascript");
  });
}

Jump to fileEdit

Add a button to go to the book file at Commons from the Index or Page namespace, and to the transcluding page from the Page namespace

mw.loader.load('//en.wikisource.org/w/index.php?title=User:Inductiveload/jump to file/load.js&action=raw&ctype=text/javascript');

MiniPaneEdit

Add a mini image page in the Page namespace for keeping the proofread text nearer the input box.

mw.loader.load('//en.wikisource.org/w/index.php?title=User:Inductiveload/MiniPane.js&action=raw&ctype=text/javascript');

IaUploadPopupEdit

Add a small popup to assist uploading works at the IA via IA-Upload

mw.loader.load('//en.wikisource.org/w/index.php?title=User:Inductiveload/IaUploadPopup.js&action=raw&ctype=text/javascript');

Scan TranscludesEdit

Show, in the pagelist grid, which pages are transcluded and which are not, colouring the pages according to expected transclusion status.

mw.loader.load("/w/index.php?title=User:Inductiveload/scan_transcludes.js&action=raw&ctype=text/javascript");

ActivePageAlertEdit

Display an icon if a page has been edited recently, with the ability to vary the definition of "recently" on a per-other-user basis.

mw.loader.load('//en.wikisource.org/w/index.php?title=User:Inductiveload/ActivePageAlert.js&action=raw&ctype=text/javascript');

TweaksEdit

Show an indicator when a script loadsEdit

I use this to check my local script is loading

$(function() {
  $(".mw-indicators").append($("<img src=\"https://upload.wikimedia.org/wikipedia/commons/thumb/c/c4/OOjs_UI_icon_chem.svg/20px-OOjs_UI_icon_chem.svg.png\">"));
});

Add nocache=1 to WS-export sidebar linksEdit

$(function() {
  $("#p-coll-print_export a[href*='ws-export.wmcloud.org/?format'").each(function(i, a) {
    $(a).attr("href", $(a).attr("href") + "&nocache=1");
  });
});

Add a button to the main page POTM to edit itEdit

if (mw.config.get("wgTitle") === "Main Page") {
  $(function() {
    $(".collaboration-potm tr:first-child td:last-child").prepend(
      $("<a>")
        .css({float: "right", "font-size": "70%"})
        .attr("href", mw.config.get("wgScript") + "?title=Module:PotM/data&action=edit")
        .text("[+]")
    )
  });
}

Add a link your your bot's contributionsEdit

$( function () {
  var botname = "InductiveBot";
  mw.util.addPortletLink(
    'p-personal',
    '/wiki/Special:Contributions/' + botname,
    botname,
    'pt-botcontribs',
    'Contributions by ' + botname,
    '',
    '#pt-logout'
  );
} );

Special icon for Internet Archive linksEdit

a.external[href^="https://archive.org"] {
    background: url("//upload.wikimedia.org/wikipedia/commons/thumb/1/13/Internet_Archive_7x8px.svg/7px-Internet_Archive_7x8px.svg.png") no-repeat right;
    /* @noflip */
    padding-right: 10px;
}

Add a custom panel of special characters to the WikiEditor toolbarEdit

	var addCharacters = function () {

		$( '#wpTextbox1' ).wikiEditor( 'addToToolbar', {
			section: 'characters',
			pages: {
				yazidi: {
					layout: 'characters',
					label: 'Yazidi',
					characters: [ 'Ḍ', 'ḍ', 'Ḳ', 'ḳ', 'Ṣ', 'ṣ', 'Ḫ', 'ḫ', 'Š', 'š', 'â', 'î' ]
				}
			}
		} );
	};

	/* Check if view is in edit mode and that the required modules are available.
 * Then, customize the toolbar … */
	if ( [ 'edit', 'submit' ].indexOf( mw.config.get( 'wgAction' ) ) !== -1 ) {
		mw.loader.using( 'user.options' ).then( function () {
		// This can be the string "0" if the user disabled the preference
		// ([[phab:T54542#555387]])
			if ( mw.user.options.get( 'usebetatoolbar' ) === 1 ) {
				$.when(
					mw.loader.using( 'ext.wikiEditor' ), $.ready
				).then( function () {
					addCharacters();
				} );
			}
		} );
	}

Show microformat data under the headerEdit

#ws-data {
    display: block !important;
}

#ws-data > span {
    margin-right: 1ex;
    padding: 0 4px;
    background-color: #e8e8e8;
    border: 1px solid #aaa;
    border-radius: 4px;
}

"Steal" the preview accesskey (p) for the "preview with this template" button insteadEdit

	$( function () {
		var sandboxPrev = $( 'input[name="wpTemplateSandboxPreview"]' );

		if ( sandboxPrev.length ) {
			var ak = 'p';
			// unbind other users of this accessKey
			$ ( '*[accessKey="' + ak + '"]' )
				.attr( 'accessKey', null );

			// add it to our button
			sandboxPrev.attr( 'accessKey', ak );
		}
	} );

One-linersEdit

PywikibotEdit

List all indexes which have a page with a linter errorEdit
python pwb.py listpages -ns:Page -linter:misnested-tag -format:"Index:{page.title}" | sed -E 's/\/[0-9]+$//' | uniq


Image ProcessingEdit

Convert all files of type X to type Y, in parallelEdit
find . -type f  -name '*.png' -print0 | parallel -0 convert {} {.}.pbm
Make an image showing only not-black and not-white pixelsEdit
convert input.png -colorspace rgb -fill white -opaque black -fill red +opaque output.png

PDF/DJVU processingEdit

Fix a PDF that chokes ImageMagic due to "bad streams"Edit
gs -o generated.pdf -sDEVICE=pdfwrite -dPDFSETTINGS=/prepress input.pdf

FunctionsEdit

Some extra functions that might be handy for other scripts:

The rant zone (aka the why you should never donate to WMF zone)Edit

 
In its rightful place, at the bottom of the list.

What does a $130 million budget not get you?

  • phab:T95878 (filed 2015) There's still no mobile editing for Wikidata, after 9 years. Seriously.
    • phab:T259183 The totally-not-beta UI is also still trash on "narrow" screens.
  • phab:T278104 Commons uploading is still totally broken for files over 100MB and no one cares, even though it's blocking IA Upload all the time and also preventing various other uploads
  • phab:T288980 WVUI is still deciding where to even put the drawing board, let alone putting anything on it
  • OOUI is now not only barely documented and horrific to actually use, but also no-one cares because it's going to get "one day, promise" replaced by WVUI
  • VisualEditor (aka the 2017 editor!) is still completely useless, and not just at Wikisource
  • Literally every data model is still work in progress for both WD and SDC and no one cares.
  • phab:T121646 (filed 2016) There's still no API for evicting data from local storage.
  • There's still no guidance for best practices for gadgets. None.
  • The chat ecosystem is completed fragmented into siloed commercial apps and trying to bridge them together is, at best, shunted to "Volunteer time". Just pay for a darn Matrix homeserver like every other FOSS project and stop pretending like everyone will use IRC in 2021 instead of leaving for Discord and Telegram. Even the WMF doesn't really use IRC, it hides its machinations on Slack so no-one can see what the hell is going on.
  • Speaking of, why does Community Tech not work in the open?

But they still need your cash for...something? And the begging banner will use every dark pattern they can fit in there.

The counter-rant zone (aka why maybe it's not all bad)Edit

  • The Special:APISandbox is brilliant
  • The Wikipedia Library is brilliant
  • Commons providing basically infinite storage is brilliant
  • Toolforge and WMCS is brilliant
  • Site Reliability knows where their towels are
  • An API that allows Pywikibot can't be all that bad

Maintenance and reportsEdit

Below are lists of pages in Wikisource which are useful for various purposes. All of these could be out of date. If you really need up-to-date reports, just leave me a note, and I will do it as soon as I can.

  • templates A list of all templates in use on enWS, along with links and usage counts.
  • wikisource pages A list of all Wikisource namespace pages.
  • portals A list of all Portal pages.
  • ws-portal redirects A list of all Wikisource pages which redirect to Portal pages. No pages should link to these.
  • ws-wp no backlink A list of Wikisource pages linking to Wikipedia pages which do not link back here.
  • false root pages A list of pages that should be subpages but aren't.
  • site-css-js: in-progress CSS tidying-up - look here for CSS and/or JS moved out of MediaWiki namespace (rather than being deleted)

Bot activitiesEdit

I operate a bot, InductiveBot, which performs minor maintenance tasks. It is based on pywikipedia and is quite flexible. If you have a specific request, please let me know on my talk page, and I'll see what I can do!

MW devEdit

Run extension linterEdit

docker-compose exec mediawiki bash -c "cd extensions/ProofreadPage && composer install"
docker-compose exec mediawiki bash -c "cd extensions/ProofreadPage && composer test"
docker-compose exec mediawiki bash -c "cd extensions/ProofreadPage && composer phan"

Run extension parser testsEdit

docker-compose exec mediawiki php tests/parser/parserTests.php --file=extensions/ProofreadPage/tests/parser/proofreadpage_pages_pagelist.txt

Or to run all the tests in a directory:

docker-compose exec mediawiki sh -c 'find extensions/ProofreadPage/tests -name "*.txt" -exec php tests/parser/parserTests.php --file={} \;'

Run extension unit testsEdit

docker-compose exec mediawiki php tests/phpunit/phpunit.php extensions/ProofreadPage/tests/phpunit

Run linterEdit

npm run-script test

LoggingEdit

  • \MWDebug::log('foobar');

Random linksEdit

mw.hook fire points
https://codesearch.wmcloud.org/search/?q=mw%5C.hook(.*)%5C.fire&i=nope&files=&excludeFiles=&repos=
Current lag
https://en.wikisource.org/w/api.php?action=query&titles=MediaWiki&format=json&maxlag=-1
All wikis edit firehose
https://event-streams.toolforge.org/
Wikisource Image Uploader uploads at Commons
https://commons.wikimedia.org/wiki/Special:RecentChanges?tagfilter=OAuth+CID%3A+2348

Useful things to shareEdit

These are some useful scripts I have hacked together. I guarantee nothing! They are certainly not always neatly coded or structured, but they work for quick and dirty jobs.

Tesseract retraining projectEdit

Very slowly, I am working on a retraining of some models for Tesseract: Tesseract

General Python scriptsEdit

GIMP scriptsEdit

  • The best one:
    • Whiten-background.scm Gimp script to remove the background of an image. An adaptation of a script by Leonid Koninin. This is fairly harsh on some images, so use with care.
  • Less good ones:
  • Remove-paper-texture.scm Gimp script to remove the paper background from a scan of a black and white image by a pretty brutal adjustment of the levels
  • Remove-background-colour.scm Gimp script to remove a flat background colour from an image. This is a fairly brutal algorithm, use with care on delicate images. Essentially, this is just a "bundle" of "select colour, erase the colour, and desaturate. It'll likely be better to use one of the above scripts, but this one is fairly instructive from a script-writing point of view.

Bibliographic junkEdit

Data module ecosystemEdit

None of these work fully. Yet.

Core modules
Client modules

One touch template wrapping with AutohotkeyEdit

If you use Autohotkey (and you should be), the following is a useful function that lets you wrap the current mouse selection in a template, which saves you having to paste in the contents.

F2 & s :: wrapTemplate("sc") ; small caps

wrapTemplate( name )  
{
    front :="{{}{{}" . name . "|"
    back :="{}}{}}"  
    wrapTags( front, back) 
    return
}

wrapTags( front, back ) 
{
    AutoTrim Off               ; Retain any leading and trailing whitespace on the clipboard.
    ClipSaved := ClipboardAll  ; Save the entire clipboard so we can restore it when we're done
    clipboard =                ; clear the clipboard
    SendInput ^x               ; cut the selection to the clipboard
    ClipWait                   ; wait for the clipboard to contain something
    SendInput %front%%clipboard%%back% ; Output what was selected, surrounded by front and back
    Clipboard := ClipSaved     ; Restore the original clipboard
    ClipSaved =                ; Free the memory in case the clipboard was very large.
    return
}

Regular expressionsEdit

Function Search pattern Replacement Pattern
Remove single newlines. Useful for OCR'd text /([^\n])\n([^\n])/g '$1 $2'
Convert relative links to static links. Useful when putting a TOC in the Page: namespace. (/\[\[\/(.*)\/\]\]/g '\[\[$1\|$1\]\]'

My requestsEdit

Works I'm keeping an eye out for.

PeriodicalsEdit

Technical wishlistEdit

Some things I'd like to see done (that isn't actual proofreading). Some of it is unimportant, some of it may be controversial and un-discussed, but would be nice to address and tighten up.

  • Get dynamic layouts to work for non scan-backed works and scrap {{prose}} and other hard-coded formatting.
  • Fix poem tags - only by having all lines as p or span-tags can we have hanging-indented continuation lines like 95% of all printed poems are. Stanzas should be divs . Might need a whole new tag in the poem extension, but might not be that hard?(???)
  • Train Tesseract specifically for 1700s-style printing esp. with long-s
  • Allow match-and-split to match to PDFs (since there are now ~1m PDFs on Commons)
  • Get para breaks working in OCR loading: phab:T230415
  • Add common fonts:
    • Cursive: phab:T166138
    • JUnicode: phab:T173573
    • Sans Outline (and remove hacks like ℕ𝔼𝕎 𝕐𝕆ℝ𝕂)
    • Serif Outline
    • Maybe a better Polytonic greek?
  • Move MediaWiki:Proofreadpage_index_template to a module
  • Move Template:Header to module

Half-doneEdit

  • Ebook review process leading to categorisation, then... Category:Ready for export
  • Improve index autofill to fetch author links from Commons creator templates: Getting there: Mediawiki:Gadget-Fill Index.js
  • Fix {{FI}} which is invoking full-size images every single time. Merge with and/or deprecate {{large image}}.
  • Fix headers on mobile, the tabular structure is unfriendly on narrow screens (main header done, other namespaces pending main header module-ificaton). See {{header/main block}}

DoneEdit