From Wikisource
Jump to navigation Jump to search
Scriptorium Scriptorium (Help) Archives, Last archive
The Scriptorium is Wikisource's community discussion page. This subpage is especially designated for requests for help from more experienced Wikisourcers. Feel free to ask questions or leave comments. You may join any current discussion or a new one. Project members can often be found in the #wikisource IRC channel (a web client is available).

This page is automatically archived by Wikisource-bot

Have you seen our help pages and FAQs?

Make a text exportable[edit]

Hello, I was wondering how to set properly a text for it to be exportable (into pdf/ePub for instance), and how to make it exportable then using templates. I haven't found so far an example of a text exportable, even looking into the texts of famous authors, while there are tons of exportable texts in the french wikisource. Thank you :) — Polochinoc (talk) 15:04, 28 May 2020 (UTC)

There is nothing in particular to do to make a text exportable: the "WS-export" tool is enabled on all pages. There should be a export link for each of ePub, Mobi and PDF in the left sidebar (or top right in mobile view). For example, this is the ePub for Agrarian Justice:
When aiming for export, avoid using certain formatting, especially fixed widths which might not fit on an ereader screen. For example, setting an element to have width:60em; will extend off the screen on most devices, prefer max-width:60em; to allow it to shrink to fit small screens. A quick check of Agrarian Justice indicates this work should be fine. Generally speaking, if it looks OK in Layout 2, it's probably be fine on an ereader.
Curating works for export is something that English WS does not do much of: there is no category or other list of "known-good" exportable works that I know of, or any procedure to check works export nicely. Inductiveloadtalk/contribs 15:32, 28 May 2020 (UTC)
There is a new project at Meta (+ accompanying discussion) aiming at export improvement. --Jan Kameníček (talk) 16:01, 28 May 2020 (UTC)
Oh, I didn't see the column at the left, thank you. Okay I understand, you don't have a category that can be added to a text, so that is it is known "good for export". By the way, I downloaded Agrarian Justice as pdf and saw it didn't keep the center alignment for the titles. But otherwise, it did render it pretty well. I'll check the pages as well — Polochinoc (talk) 16:12, 28 May 2020 (UTC)
Hmm, yes,the PDF using the side-bar link does lose the centre (and right) formatting, and seems to use a different for the tables too. The sidebar actually uses the "ElectronPdf" tool. Ws-export also can do PDFs, which does seem get the centre/right formatting right: Hopefully the project Jan linked above will address some of these quirks. I'd say that for me, generally, ePub and Mobi are the critical targets, since they allow the device to reflow text as needed for the small screen. Inductiveloadtalk/contribs 18:04, 28 May 2020 (UTC)
@Polochinoc, @Inductiveload: I think the Tech Team will appreciate if you write them what you observe that needs to be improved. --Jan Kameníček (talk) 07:13, 29 May 2020 (UTC)
I think, for a large proportion of enWS issues, problems are local:
  • Their first example of misalignment was Page:Trees Kilmer.djvu/9, which was using a <center> tag to centre a bordered div. This should be CSS margin:auto;. Possibly, WS-export could fix this on the fly, but it's better to just use the right markup in the first place. Agrarian Justice had this too with the tables, which used {| align=center instead of {| style="margin:auto;".
  • The same work (Trees and Other Poems) also does not use {{page break}}, so the ebook doesn't contain hard page breaks in the right place, so it has to split the div across pages as it doesn't know better. I don't think this is a defect in the WS-export, it's just how ePub/ereaders work when there's a div that won't fit on a page.
  • Works that use wide fixed-width formatting are likewise going to break ebook formatting through no particular fault of the export layer.
  • ElectronPDF should IMO, just, not be used at WS. It doesn't handle multipart works, so it's basically useless for any work with subpages. Agrarian Justice was just lucky in that it doesn't have subpages. Why not just use the WS-export PDF engine, which handles subpages and at least will be consistent with the eBub/Mobi outputs in terms of what pages are included? Plus ElectronPDF seems to strip quite a lot of a formatting, which probably makes sense when printing Wikipedia articles, but not so much for WS. What appears in the left sidebar is at least partly under local enWS control: WS-export options are from MediaWiki:Gadget-WSexport.js. I'm unsure ElectronPDF even maintained: mw:PDF_export says it's deprecated.
  • Making exported works first-class citizens like frWS (which has very high ebook engagement rates) does with clear icons in headers, prominent icons on all front-page works and categorisation into Category:Bon pour export is again down to local policy and not tool defects.
  • Having such a category would enable us to present an OPDS catalog of exportable works, so WS can be integrated directly into ebook readers like MoonReader+ and managers Calibre as a "net library". As it stands, an ODPS catalog of enWS works is likely to present rather variable-quality results, as not all works have been made with export in mind.
  • On-wiki instructions for eBooks are scattered and hugely dated: Wikisource:eBook and Help:Reading offline. Our first suggestion for finding an ebook is to find one for sale!
I'll try to distil some examples that I think are tooling-related and report them to the project. Inductiveloadtalk/contribs 09:14, 29 May 2020 (UTC)
@Inductiveload: Regarding Agrarian Justice{| align=center is wikimarkup, not HTML, so it's not really the equivalent of using <center></center> (in 2020?!?): it's a "bug" (fsvo) in Mediawiki more than a local issue. Not that we can't or shouldn't work around it for the benefit of our (ebook) users, but fair's fair.
Where is it you think there should be a {{page break}} that there isn't? The work does use them in what to me appears roughly the standard way, and the ePub looks fine here.
And generally, if there are significant local issues at enWS then it would seem a good first step would be to document them somewhere so we can learn and adapt. A list of observed problems, and a set of derived "Do's and Dont's", would be most helpful for those of us who almost never look at an ePub version. And it may show opportunity for making templates and other tools "ePub friendly" by default where they aren't today. --Xover (talk) 10:14, 29 May 2020 (UTC)
  • Sure, I meant alignment issues in general: center tags and align=center both don't work in ebooks. It's really not the ebook layer's fault that that alignment failed in either case. Perhaps it could be intercepted and tweaked by WS-export, but we can also just use margin:auto locally and it's done. I made a few changes to Agrarian Justice to tidy up the ePub (the only real issue was the table alignment), and it exports well (suitable for Category:Good for export, I say!).
  • Re page breaks: I meant in Trees and Other Poems (which is indefinitely protected): probably in the main namespace, between the "non-flowed" front matter pages. This would (probably?) resolve the issue reported in the meta project: meta:Community_Tech/Ebook_Export_Improvement#Formatting_&_Styles (the alignment failure being caused by a center tag, now resolved).
  • I did recently propose a list of suggestions for for making our works export-friendly, but didn't get support because a very small number of works have formatting that's not easy to ebook-ify: Wikisource:Scriptorium/Archives/2020-01#Curating_works_for_export. No-one seemed to support (or oppose for that matter) the idea of having a frWS-style "known good for export" process of some sort. I'd be more than happy to start the discussion again.
  • I have collected some material on some things for consideration when formatting ebooks: User:Inductiveload/Sandbox/Formatting for export Inductiveloadtalk/contribs 10:36, 29 May 2020 (UTC)
@Clockery, @Londonjackbooks, @EncycloPetey, @BethNaught, @Beleg Tâl: Note that in order to address the ebook problem flagged above I have modified the (featured) text at Trees and Other Poems to add {{pb}} between distinct pages of the front matter (particularly those that feature non-reflowable content, like the title page). Please let me know if you have any concerns or objections to this.
@Inductiveload: I think a quality control process that ends up with verified good ebook exports tagged with Category:Good for export is a good idea, at the very least so long as there is a significant chance that our works are not good for ebook export by default. Perhaps such checks should be incorporated into the featured text process too? --Xover (talk) 14:58, 29 May 2020 (UTC)
Thanks! The different front matter pages now render in my ebook readers (MoonReader+ and Calibre) on separate pages. The problem noted at meta no longer occurs (unless your font size is set so large and your screen so small that the title page content simply doesn't fit on the page at all, but there's not a lot to be done about that).
Re quality control process, I think there are a two parallel aims here:
  • Have some kind of process to check for the basic gotcha's for ebooks so people can check off works. I would say that egregious ebook fails are not that common, most works come out pretty well, all things considered. I can't see any critical issues on a selection of the current recent texts, for example. However, I would certainly says it's common that works can benefit from small tweaks to improve the ebook output, even after reaching validated status. Generally speaking, the tweaks are minor and don't affect the "normal" view of the work. For example, changing a "width:30em" to "max-width:30em" doesn't have any effect until the screen is narrower than 30em. I think it's rare indeed that a work is outright unsuitable for an ebook.
  • Some kind of listing mechanism, e.g. a category. Categories are good because they're easy to maintain (no manual formatting), well understood, easy to reason about and, critically, easy to query automatically, e.g. for providing a library catalogue of verified ebooks. We have Validated texts and Proofread texts, but most pages there are subpages or article pages (e.g. DNB00) which will clog up the list. IMO, "good for export" entails more than just proofreading status and should also be restricted to "top-level" works that you'd expect to see in such a catalogue.
I think the most critical thing about the process is that it should 1) be easy, to encourage people to do it and expand our repertoire of verified-nice ebooks and 2) should not require onerous or invasive changes to works, even if that means some works don't ever get a "good for export" tag. Inductiveloadtalk/contribs 15:35, 29 May 2020 (UTC)
@Xover: I always use page breaks (generally {{ppb}}) in front matter, I think there are very few works which would not be improved by inserting these in the appropriate places. —Beleg Tâl (talk) 03:15, 11 June 2020 (UTC)
+1 and have been for years+for+ages. Where the series it is not flowing the separation of the idea is best IMNSHO emphasised with a break. — billinghurst sDrewth 05:14, 11 June 2020 (UTC)

A question about pdf scans[edit]

In the page Help:Adding texts, it’s written that before typing the text in Wikisource, I must first upload the scan to wikimedia commons. There’s still one thing that I don’t understand though, I checked through other source like Republic of South Africa Constitution Act, 1961 and Unilateral Declaration of Independence, and I couldn’t find anywhere where they linked the scan. Do I just upload the scan to wikimedia commons without linking it anywhere on Wikisource? Also, if we enter wikimedia commons, we see that it’s not used anywhere in Wikisource. Thanks in advance, -RiverThames27 (talk) 11:43, 31 May 2020 (UTC)

Uploading scans is still not compulsory but it is true that it is highly preferred. The works which you have linked to were added to English Wikisource long time ago (in 2006 and 2009), i. e. before the pracice of uploading scans started. For an example of a scan-backed document see e. g. The Constitution of the Czechoslovak Republic.
If you want to add a scan-backed work, you have to follow these steps:
  1. Upload the scan to Commons (if it is in public domain both in the United States and its source country) or here (if it is public domain only in the United States).
  2. Found an index page (example).
  3. Profread the individual pages
  4. Transclude the work into the main namespace.
These steps are generelly not too dificult, but they may seem difficult to a novice contributor. For these reasons do not hesitate to aks for help at any stage of the work, you will always find help here. If you upload the scan, I can found the index page for you if needed. --Jan Kameníček (talk) 13:51, 31 May 2020 (UTC)
Thanks for the reply! I still have one question though. The text I would like to add is the 1965 constitution of Rhodesia, and it can be found here: It should be public domain in the U.S. because it’s a government document, but it’s source country doesn’t exist anymore. Should I upload it to Wikisource or to wikimedia commons? If to wikimedia commons, should the index page be in Wikisource or in wikimedia commons, thanks in advance, -RiverThames27 (talk) 09:34, 1 June 2020 (UTC)
Oh, and one more question, in this source the chapters don’t always begin on a new page. How will the Transclusion work? I was recommended to search on splitting it, which I haven’t done yet, but even then you have chapters that start in the middle of the page. Thanks in advance, RiverThames27 (talk) 09:38, 1 June 2020 (UTC)
Hello? -RiverThames27 (talk) 14:20, 2 June 2020 (UTC)
@RiverThames27: Unfortunately, I am not able to answer the question about Rhodesia. Maybe @Xover: or @Prosfilaes: could help.
As for chapters which do not begin on a new page: in such a case you must devide the page into sections and then transclude only the particular section.
The simplest way to divide the page is to use:
## s1 ##
Text of section one
## s2 ##
Text of section two
To transclude the text upto some section or from some section you use for example
<pages index="name of book.djvu" from=6 to=9 fromsection=s2 tosection=s1 />
<pages index="name of book.djvu" include=6 onlysection=s2 />
If needed, I can do the transclusion after the text is proofread. --Jan Kameníček (talk) 15:40, 2 June 2020 (UTC)
Thank you very much! -RiverThames27 (talk) 15:45, 2 June 2020 (UTC)
Oh and btw, I searched for ways to split the pdf pages since it's scanned in such a way that it shows 2 pages on 1 page, and all I have found was on extracting individual pages, and the only one that actually was what I was seacrhing for required a payed version of Acrobat Reader DC which I simply don't have. Since new chapters in any case begin in the middle of the page, would it be acceptable to upload this one despite the fact that one page actually contains several pages? And I would prefer to upload it to wikimedia commons so that it can be referred to in other wikis as well, but if ends up being copyrighted in it's country of origion (which doesn't exist anymore) I guess I will have to uplaod to wikisource. Thanks in advance, -RiverThames27 (talk) 16:37, 2 June 2020 (UTC)
I am very sorry, that is another thing I am not able to help with. I know Xover often helps people with scans, but according to the information on his user page he has been quite busy recently. But you can try to ask him on his talk page where he is more likely to answer. --Jan Kameníček (talk) 17:08, 2 June 2020 (UTC)
I've used gscan2pdf to split the side-by-side pages of the PDF. I uploaded the result to Commons and made a quick index at Index:Constitution of Rhodesia, 1965.pdf. Based on the wording that includes "antecedents of Zimbabwe" in the Commons commons:Template:PD-Zimbabwe template, I think it'll be OK at Commons. The file lacks an OCR text layer, but I'm not set up to deal with that at the moment. In the meantime, there should be an OCR button in your toolbar (which seems not to work), and there's a Google OCR button that you can enable here: Special:Preferences#mw-prefsection-gadgets. Inductiveloadtalk/contribs 17:09, 2 June 2020 (UTC)
Thank you very much! -RiverThames27 (talk) 18:14, 2 June 2020 (UTC)
@RiverThames27: On the copyright situation… Rhodesia was strictly speaking a British colony at the time this work was published, which would mean a Crown Copyright term of 50 years from publication. However, I would be inclined to say Zimbabwe is the successor in interest and the governing law; and in Zimbabwe the text of laws and related material is not eligible for copyright. The same is true in the US. That is, worst case (British law controlling), the copyright expired in 1965 + 50 = 2015. In other words, Commons should be fine. Oh, and the Index: is always here on Wikisource; we just outsource hosting the scan files themselves to Commons.
@Inductiveload: I'm tooled up for OCR with DjVu files, so I made a .djvu with text layer at File:Constitution of Rhodesia, 1965.djvu. I didn't bother making a separate index for it (if you prefer that one we can move the index and pages over; but if the Google OCR results are good enough it's probably not worth the effort).
@Jan.Kamenicek: Thanks for the ping. Right now my time is more unpredictable than all that limited, but I'm leaving the notice up because that situation will change based on the IRL situation in my neck of the woods. --Xover (talk) 18:59, 2 June 2020 (UTC)
@Xover: Thanks for the reply! About the OCR, I have already started typing one of the pages manually. Should I now just let it do that automatically? Thanks in advance, -RiverThames27 (talk) 19:12, 2 June 2020 (UTC)
@RiverThames27: That's up to you: the point of OCR is to save you manual typing.
PDF and DjVu-format files can contain a hidden text layer that Mediawiki knows to extract. When the file has such a text layer you will find the text box for each page prepopulated with that text. In addition, we have gadgets that give you an OCR button in the toolbar on each page. When you press the button the page image is sent to a server that runs OCR on it and sends back the results. In either case you end up with some text in the text field, it's just that in the latter case you had to press an extra button to get it.
The PDF file that you're currently transcribing against doesn't have a hidden text layer, so you'll need to use the OCR button. If you're happy with the results it gives you then there's no real reason to do anything else. If the results are very poor that way, we can try to move the Index: and Page: pages over to use the DjVu file which does have a text layer (that was made with a different OCR engine). It can some times make a big difference, but most of the time it's just six with one and half a dozen with the other. There's usually not a big reason to do so, I just happened to already have it processing before I saw that Inductiveload had already uploaded a PDF.
And if you want to type everything manually that's of course fine too. --Xover (talk) 19:56, 2 June 2020 (UTC)
Great! Of course, I'll be happy for the existing index to be migrated to the DjVu. Google OCR seems pretty good, but it's friendly to have the text layer. Inductiveloadtalk/contribs 21:29, 2 June 2020 (UTC)
@Xover, @Inductiveload: About migrating to the DjVu, is it just moving the page with Special:MovePage/Index:Constitution of Rhodesia, 1965.pdf to Index:Constitution of Rhodesia, 1965.djvu, and then repeat the same with the other pages that already exist, or is there more to it? Thanks in advance, -RiverThames27 (talk) 10:23, 3 June 2020 (UTC)
For this text it’s necessary since when I tried the OCR it recognised for the top part just the numbers, and for the bottom part just the text (it was a page from the table of contents) -RiverThames27 (talk) 10:25, 3 June 2020 (UTC)
It's better for an admin to do it, otherwise it makes a bit of a mess of redirects. For the OCR, try the Google OCR button, it's works well for the contents pages in this work. You can enable it under "Gadgets" in your preferences (look near the bottom). Inductiveloadtalk/contribs 10:34, 3 June 2020 (UTC)
I used the google OCR button and it’s actually quite good. It didn’t put it in the templates so it was all in one line at the beginning, but I have sorted this out. Thanks @Inductiveload:! -RiverThames27 (talk) 12:11, 3 June 2020 (UTC)
@RiverThames27: No OCR is going to give you a finished page; it'll always be just the text and with various levels of accuracy in recognising the letters. For some works the OCR engine used can make a big difference because different engines have different strengths and weaknesses. For most works it makes little difference. If Google OCR gives reasonable results you'll probably want to stick with that. For comparison, you can trick Mediawiki into showing you the hidden text layer of a file by going to a Page: link even if it doesn't have an index. So here are page 4 and 5 from the DjVu file: 4, 5 (please don't create these pages; they're not connected to anything). As you can see these need a lot of manual work too. If you you still want to switch to the DjVu then just let me know and I'll move the Index: and Page: pages over for you. --Xover (talk) 13:48, 3 June 2020 (UTC)
@Xover: I think the pdf version is good enough, google ocr is quite good in detecting text and I think the things that I actually had to correct were minor. -RiverThames27 (talk) 15:02, 3 June 2020 (UTC)
@Xover, @Inductiveload: Now all the table of contents is done. However, it shows me “This text needs to be proofread” How do I change it to “This text has been proofread but needs validating?” Thanks in advance, -RiverThames27 (talk) 15:24, 4 June 2020 (UTC)
Like in pages 1,6, and 7 -RiverThames27 (talk) 15:29, 4 June 2020 (UTC)
There are some buttons under the edit window to set the page proofread status. See Help:Page status for the whole process.

By the way, I have set up a template, {{CoR65 para/s}} to assist with the formatting, as those hanging paragraphs can be a bit fiddly. You can check the first few pages of the main text to see how they can be used. Inductiveloadtalk/contribs 15:44, 4 June 2020 (UTC)

Thank you very much! And btw, what’s with the page numbering? Why is there one more page 43 after page 92? I know it links to the proper page, but why does the text say 43? Thanks in advance, -RiverThames27 (talk) 17:20, 4 June 2020 (UTC)
That was in the original document. It is apparently the relevant clipping from the Southern Rhodesia Government Gazette (actually a Supplement). The page numbers for those pages are "1315 (43)" and "1315 (44)", where 1315 is (probably) the issue of the Gazette and 43/4 is the page number. Inductiveloadtalk/contribs 17:38, 4 June 2020 (UTC)
Thanks for you reply! Should those numbers stay that way or change? And can you check if I did the page Page:Constitution of Rhodesia, 1965.pdf/12 properly? I tried to imitate what you did in Page:Constitution of Rhodesia, 1965.pdf/11, but when I did the same that you did with the a, b, c it put the text after that in boxes like pre-formatted text. Thanks in advance, -RiverThames27 (talk) 17:48, 4 June 2020 (UTC)
You had a couple of misaligned start (/s) and end (/e) templates. There has to be exactly one /s for each /e. I fixed up p12 and now it seems to work OK. I also added the sidenotes. Inductiveloadtalk/contribs 18:03, 4 June 2020 (UTC)
Thank you very much! -RiverThames27 (talk) 18:46, 4 June 2020 (UTC)
Btw, I started working on Page:Constitution of Rhodesia, 1965.pdf/13, and every next line starts from the very beginning of the line. How did you make it so that every line starts from the same place as the previous one? Thanks in advance, -RiverThames27 (talk) 19:27, 4 June 2020 (UTC)
And btw, why does the progress say “Source file needs an OCR tex layer” if some of the pages are already ready? Thanks in advance, -RiverThames27 (talk) 19:32, 4 June 2020 (UTC)
You have to change the first template parameter from "0" (no level) to 1, 2 or 3 depending on the depth of the outermost number.
Progress is set manually with a drop-down in the Index page when you edit it. This file has no OCR text in it (you have to use the OCR button; if it had OCR, the OCR text would load automatically when you create the page). It doesn't mean you can't proofread it. When all pages are yellow, you can change the drop-down to "Needs validation". Inductiveloadtalk/contribs 20:09, 4 June 2020 (UTC)
Thank you very much! Now it works! Is there a way to make the side notes actually on the side like in the original text? Some of them are left side notes and others are right but it doesn’t seem to make any difference. Thanks in advance, -RiverThames27 (talk) 08:51, 5 June 2020 (UTC)
Yes, change {{left sidenote}} to {{right sidenote}}. Note: when "trancluded" (displayed in the main namespace), they will be rendered differently, according the the layout. Don't worry two much about perfect formatting for sidenotes, they will always be slightly different on a computer than they are on paper. Inductiveloadtalk/contribs 09:07, 5 June 2020 (UTC)
I mean some of them are left sidenote and others are right sidenote. They seem to look the same in practice. What’s the difference? -RiverThames27 (talk) 15:38, 5 June 2020 (UTC)
In the page Page:Constitution of Rhodesia, 1965.pdf/13 The first sidenote is a left sidenote and the second one is a right sidenote. What’s the difference? They look the same to me. Thanks in advance, -RiverThames27 (talk) 15:40, 5 June 2020 (UTC)
@Inductiveload, @Xover: What’s the difference between a left sidenote and a right sidenote because both appear exactly the same on screen. Thanks in advance-RiverThames27 (talk) 12:09, 6 June 2020 (UTC)
@Inductiveload, @Xover: Hello? -RiverThames27 (talk) 16:41, 6 June 2020 (UTC)
@RiverThames27: Left and right sidenotes display on the left and right side of the page, respectively, in the Page: namespace. When we transclude it for presentation in the main namespace the dynamic layouts will kick in and reformat them to fit the active layout. Looking at Page:Constitution of Rhodesia, 1965.pdf/13 the two first sidenotes are indeed on the left and then right side. The second one is hard to see on my computer due to the really rather unfortunate layout problems of sidenotes in the Page: namespace, but it definitely is there and displayed to the right. Does this not happen on your computer? --Xover (talk) 07:06, 7 June 2020 (UTC)
@Inductiveload: I was editing on my ipad, so it showed all the sidenotes on top of the text. I opened it on my computer, and it's a mess! It shows the text on the left part of the screen, and the image on the right part, and they collide with each other. Also, I noticed that there is a difference between the left sidenote and the right sidenote (on my ipad, they looked the same), and when I changed the right sidenote to a right sidenote, it started overlapping with the sidenote that was above! Is there any way to show the image bellow the text? Thanks in advance, -RiverThames27 (talk) 11:30, 7 June 2020 (UTC)
@Xover: Sorry, for some reason I didn’t notice you left the message and typed ping Inductiveload😅 -RiverThames27 (talk) 13:52, 7 June 2020 (UTC)
@RiverThames27: It is a mess in the Page: namespace, sorry, nothing to be done about that. It'll look better once it gets transcluded for presentation. The problem is that Mediawiki doesn't really have support for sidenotes so we've had to create a kinda hacky workaround. You can see an example of how sidenotes will look on The Solar System/Chapter 1.
There are also buttons in the editing toolbar to switch the order of the text edit field and the scan image (in the "Proofread tools" section; looks kinda like a Tetris block or something), but in my experience only the default side-by-side layout works well in practice. --Xover (talk) 17:40, 7 June 2020 (UTC)
However, it would be nice if at least the line-height parameter were allowed to be set up manually. If the line-height were smaller than the default 140%, the sidenotes would interfere much less.--Jan Kameníček (talk) 17:50, 7 June 2020 (UTC)
@Xover, @Jan.Kamenicek: Thanks for the reply! I understand. Surely there’s someplace where it’s possible to request support for sidenotes, isn’t there? In meta or in Mediawiki or somewhere. -RiverThames27 (talk) 19:42, 7 June 2020 (UTC)
Specific changes to the templates like {{Left sidenote}} can be asked and discussed here. As for changes to Mediawiki so that it supported sidenotes better, you can try Phabricator, but my experience with their responsiveness is not very good. --Jan Kameníček (talk) 20:01, 7 June 2020 (UTC)


@Jan.Kamenicek: setting line-height for sidenotes would require them to be divs, not spans, which would probably cause issues when they occur inside paragraphs. As they are (spans), they can't have a smaller line-height than the surrounding text. You can set it smaller in CSS, but it won't actually get smaller. It's the same reason {{smaller}} doesn't make line-height smaller, but {{smaller block}} does.
Adding display: inline-block to the span might be enough.
@RiverThames27: it's not really something that could be added from the MediaWiki side, it's more a problem finding a way to display then using HTML and CSS that works on all devices and screen sizes. Not a lot of love has been shown to them in the Page namespace, because that's a working area, the "presentation" area is in the Main namespace, where it's quite a bit better (but still not perfect).
There is a bit of scope in changing {{sidenotes begin}} for turning off one side, which would decompress the pages a bit, and then making the main column shrink to fit. Small windows will still be very tight, but might not overlap so terribly. Inductiveloadtalk/contribs 20:08, 7 June 2020 (UTC)
@Inductiveload: Well, I guess it will look fine🤷🏻‍♂️. It’s just weird that they the image shows up on the right side of the page, don’t they know that the text overlaps it? -RiverThames27 (talk) 12:12, 8 June 2020 (UTC)
"They" (by which I assume you mean the MediaWiki developers) probably don't really know, because we at enWS caused it though setting fixed widths (35em + 11em * 2) in the local {{sidenotes begin}} template. There's not a lot MW can do to help us that couldn't be sorted with a overhaul of the templates here at enWS first. Sadly, it'll always be a bit tricky as fitting six text columns (four sidenotes, two each for preview and scan, and two text bodies) size-by-side is always going to be a squeeze. Inductiveloadtalk/contribs 12:31, 8 June 2020 (UTC)
@Inductiveload: Oh. I see. Isn’t there an option to move the image though so that the text won’t overlap with the image? Thanks in advance, -RiverThames27 (talk) 12:38, 8 June 2020 (UTC)
There is an option to have the page vertically stacked in Edit mode (look in your preferences for the permanent setting, or there's a button in the edit toolbar to change temporarily). If you really want to have the stacked interface even when not editing, you could add the following to User:RiverThames27/common.css:
 * Top/bottom layout in Page namespace
.prp-page-content {
	float: none;
.prp-page-image {
	margin: 0 auto;
	float: none;
This is an ugly hack, but it might be good enough for your purposes. The real solutions are: 1) use a higher-resolution monitor (i.e. blame you, the victim) and, better, 2) we fix our templates. Inductiveloadtalk/contribs 12:52, 8 June 2020 (UTC)
For anyone with interest in sidenotes: I have a proposal at Template_talk:Sidenotes_begin#New_approach for a change that would prevent sidenotes spilling over the page image in Page namespace. If my calculations are correct, it should be a drop-in replacement in terms of parameters. It's not a resolution to every single sidenote issue, but it's a start and doesn't require botting all 21786 users of {{sidenotes begin}}. Inductiveloadtalk/contribs 14:38, 8 June 2020 (UTC)
@Inductiveload: Sounds good, but I’m not really qualified to make such decisions. -RiverThames27 (talk) 09:23, 9 June 2020 (UTC)
@Inductiveload, @Xover, @Jan.Kamenicek: Oh btw, some of the pages end with CoR65 para/e despite the fact that the paragraph continues on the next page. Will that make the text in the finished product start in a new line despite the fact that it’s the same paragraph? Thanks in advance, -RiverThames27 (talk) 20:03, 12 June 2020 (UTC)
An example: Page:Constitution of Rhodesia, 1965.pdf/14 -RiverThames27 (talk) 20:05, 12 June 2020 (UTC)
Put the end template in the footer, and the start template on the next page in the header and it should work. The header and footer are not transcluded into the main namespace. See Page:Constitution of Rhodesia, 1965.pdf/11 for example. Inductiveloadtalk/contribs 20:48, 12 June 2020 (UTC)
@Inductiveload: Thank you very much! I will fix it right away. -RiverThames27 (talk) 09:32, 13 June 2020 (UTC)
I put the end template on the footer just on pages where the the paragraph didn’t end and the start template on the header just on pages where the the first paragraph is actually the continuation of the paragraph on the previous page. That’s what should be done, right? And btw, what to do with pages where a word continues on the next page? For example, the word subscribe starts at the end of Page:Constitution of Rhodesia, 1965.pdf/15 and ends the beginning of Page:Constitution of Rhodesia, 1965.pdf/16. Will that effect the way it looks in the end product? I deleted the - so that it won’t show up in the end product, but they show up with a space in between? Thanks in advance, -RiverThames27 (talk) 09:51, 13 June 2020 (UTC)
@RiverThames27: If a word is split between pages, just leave it as sub- and scribe, and Proofread Page will remove the hyphen and join the words without a space character when transcluded. If you need to preserve the hyphen of a hyphenated word (like off-campus), replace the hyphen with the template {{peh}}. You can, if you prefer, get fancy and use {{hws}}/{{hwe}}, but I don't generally recommend that (it used to be needed for all split words, but these days it's only when a word is split inside a footnote, and it is needless complexity). --Xover (talk) 11:08, 13 June 2020 (UTC)
@Xover: Thank you very much! I added the hyphen back. -RiverThames27 (talk) 11:48, 13 June 2020 (UTC)
@Inductiveload: While editing the pagePage:Constitution of Rhodesia, 1965.pdf/20, I run into a problem with putting the title. Actually, I forgot how to do this😅, so I started searching in other pages how to do this. Then I noticed that in the page Page:Constitution of Rhodesia, 1965.pdf/13 there’s a problem with the chapters. The text Chapter III, and the title of chapter III appear, but the text part 1-composition doesn’t. Then I tried to search in the table of contents how to fix the chapter problem, and just made it worse, so I deleted it all together. However, I noticed that in Page:Constitution of Rhodesia, 1965.pdf/10 the text Chapter II and the title appear on the same line, which is also a bug that I don’t know how to fix. Can you please check what’s the problem? Thanks in advance, -RiverThames27 (talk) 12:50, 16 June 2020 (UTC)
You tried to use a template that only works inside the TOC table outside of a table. Just format it with {{center}} as normal text. For page /10, I just added a line after the center template opened. Inductiveloadtalk/contribs 12:57, 16 June 2020 (UTC)
Thank you very much! I wonder if there’s a page for the longest discussions in ws:s/h🤔 Since this discussion is quite long😅. -RiverThames27 (talk) 13:01, 16 June 2020 (UTC)
@Inductiveload, @Xover, @Jan.Kamenicek: Btw, about chapters that start in the middle of the page, I planned to leave this question for after all the pages are proofread, but it would be a pain to go through all the pages and fix it. I need to put the #s with s1 on the very top of the page, and them the #s with s2 just above the text Chapter <Chapter number>, right? And I only need to do this in pages where there’s a beginning of a chapter, there’s no point putting the #s with s1 on every page, right? Thanks in advance, -RiverThames27 (talk) 09:19, 17 June 2020 (UTC)
Correct on both counts. Later, you can refer to H:Transclusion#sections, which explains how to place only that section on the final page. Inductiveloadtalk/contribs 09:24, 17 June 2020 (UTC)
Thank you very much! -RiverThames27 (talk) 19:44, 17 June 2020 (UTC)
@Inductiveload: I tried to follow these instructions with the following page: Page:Constitution of Rhodesia, 1965.pdf/9 (the last page of the table of contents, and the first page of the actual constitution, but it just shows: 1. 1. "s1" ##. Could you please check what’s wrong? And also, should I put the label for section 2 before or after the text “IT is hereby notified that we, the Government of Rhodesia, have adopted, enacted and given to Rhodesia this Constitution.” I put it after and right before the text chapter one. And I also accidentally made it so that the text chapter I appears on the same line that the title. Thanks in advance, -RiverThames27 (talk) 19:55, 17 June 2020 (UTC)
The problem is caused by the quotation marks in ## "s1" ##, it should be just ## s1 ## --Jan Kameníček (talk) 20:12, 17 June 2020 (UTC)
Btw, I have managed to fix the problem with the title. -RiverThames27 (talk) 20:00, 17 June 2020 (UTC)
1) Don't use quotes in the section name, it breaks the magic. 2) It needs a {{nop}} to prevent the table row merging with the section header. 3) It also needed a nop at the top of each page to stop the first row of each page merging with the last row of the prior page. Yes, it's weird, sorry! It now transcludes OK with <pages index="Constitution of Rhodesia, 1965.pdf" from=1 to=9 tosection=s1 />.
@Inductiveload, @Jan.Kamenicek: What is the meaning of the section dummy in Page:Constitution of Rhodesia, 1965.pdf/9. Is it a must to put it in between every two chapters or is it something specific here because it’s in the table of contents? Thanks in advance, -RiverThames27 (talk) 09:23, 24 June 2020 (UTC)
All that means is the {{dhr}} can be excluded when the page is transcluded (<page .... to=9 tosection="s1"/> and <page .... from=9 fromsection="s2"/>). You could equally not use a dummy section and have it at the end of "s1" and no-one would notice, it'd just make a tiny blank gap at the end of the transcluded content. There is nothing special about the word "dummy", it's just another section. Inductiveloadtalk/contribs 09:57, 24 June 2020 (UTC)
@Inductiveload: Thank you very much! -RiverThames27 (talk) 20:29, 24 June 2020 (UTC)

{{Dotted TOC page listing}} formatting query[edit]

Is it possible with {{Dotted TOC page listing}} to have the page numbers nestling in under the entry column - as here, rather than out in Col3? Thanks. CharlesSpencer (talk) 17:52, 9 June 2020 (UTC)

@CharlesSpencer: Unforturnately it is not, somebody would have to either rewrite the template to add this possibility or design a different template. --Jan Kameníček (talk) 06:41, 12 June 2020 (UTC)
Thanks! CharlesSpencer (talk) 11:22, 12 June 2020 (UTC)

Header displayed twice in DNB page[edit]

I was visiting the Ponsonby, Henry (DNB00) page and see the header being displayed twice. Tried to fix it by removing {{DNB00}}, but in the preview it appears that removing the template results in the header not being displayed at all. Since I am not familiar with how the pages tag works, appreciate if anyone could lend a hand fixing it. Thanks in advance. ネイ (talk) 16:06, 10 June 2020 (UTC)

@ネイ, @ネイ: I'm trying to track it down. Will follow up on WS:S (there's an existing thread there). --Xover (talk) 17:12, 10 June 2020 (UTC)
Thanks! I'll have a look at that thread. ネイ (talk) 17:15, 10 June 2020 (UTC)


Is there a guide on TOC? ⰕⰑⰞⰀ 03:38, 12 June 2020 (UTC)

I am afraid there is not. It is preferred when the work including the TOC pages is scan-backed, proofread in the Page namespace and then transcluded into the main namespace. And because the TOC is usually different in different books, it is not prescribed what it should look like in any guide here. --Jan Kameníček (talk) 06:38, 12 June 2020 (UTC)


I am recently planning to work on Index:National Anthem Ordinance.pdf, and I encountered some sheet music inside of the related file. Of course it will be easy to use CropTool on Commons. But how should I crop out images from a local English Wikisource file? Are there any tools on English Wikisource to deal with this? Many thanks.廣九直通車 (talk) 11:49, 14 June 2020 (UTC)

We don't have any image manipulation tools, as we tend to not host images. We leve the tools and processes at Commons. — billinghurst sDrewth 13:08, 14 June 2020 (UTC)
@廣九直通車: I extracted these with the "pdfimages" command: File:National Anthem Ordinance - Stave - English.png and File:National Anthem Ordinance - Numbered - English.png. At least the first one should be possible to use the Lilypond extension for, but I have not yet learned how to do that. Please check the image metadata is correct. Inductiveloadtalk/contribs 13:29, 14 June 2020 (UTC)
@Inductiveload:Thanks! It seems that there are no problems. And also, as the melody of March of the Volunteers fulfills the Commons requirement (c:Template:PD-China and Template:Pd/1996), I'll soon send them to Commons.Nope, seems that what I'm originally thinking is wrong.廣九直通車 (talk) 14:10, 14 June 2020 (UTC)

Archives of this page and main Scriptorium[edit]

[Moved from huge unrelated section] Inductiveloadtalk/contribs 20:22, 17 June 2020 (UTC)

And a question that isn’t really related to this discussion, why does the page wikisource:Scriptorium/Help/Archives only have discussions till 2015? Where will this discussion fro example go to after it gets archived? And also they probably wanted to put there a picture, but it says File:....... -RiverThames27 (talk) 20:05, 17 June 2020 (UTC)
About the archives, I found that you can in any case enter wikisource:Scriptorium/Help/Archives/2016 or wikisource:Scriptorium/Help/Archives/2018, just for some reason they aren’t linked in wikisource:Scriptorium/Help/Archives. -RiverThames27 (talk) 20:09, 17 June 2020 (UTC)
I suggest to start this point under a new heading at the bottom, as this is already getting very long. --Jan Kameníček (talk) 20:12, 17 June 2020 (UTC)
That's a very good question. I have no idea. Looks like the index bot has gone AWOL. Archives are at Wikisource:Scriptorium/Help/Archives/2020, etc. I fixed the bad formatting. Inductiveloadtalk/contribs 20:22, 17 June 2020 (UTC)

Transclusion issue[edit]

I'm having an issue – can anybody tell me what I'm missing that's stopping the last source page from transcluding here? I can't see what's going wrong here. It seemed to work before, but doesn't any longer. I also had an issue where the templates for the footnotes and CC-by-4.0 stopped working, but this seems to have fixed itself now. Any ideas?— 🐗 Griceylipper (✉️) 05:28, 20 June 2020 (UTC)

Yep, the templating is blowing the expansion limit. (open the source to see this report at the base)
NewPP limit report
Parsed by mw1401
Cached time: 20200620065226
Cache expiry: 2592000
Dynamic content: false
Complications: [vary‐page‐id]
CPU time usage: 4.040 seconds
Real time usage: 4.110 seconds
Preprocessor visited node count: 538523/1000000
Post‐expand include size: 2093245/2097152 bytes
Template argument size: 95400/2097152 bytes
Highest expansion depth: 11/40
Expensive parser function count: 2/500
Unstrip recursion depth: 1/20
Unstrip post‐expand size: 178994/5000000 bytes
Number of Wikibase entities loaded: 0/400
Lua time usage: 0.008/10.000 seconds
Lua memory usage: 726 KB/50 MB

Transclusion expansion time report (%,ms,calls,template)
100.00% 4015.360      1 -total
 89.25% 3583.772   1625 Template:Nornabr
 83.92% 3369.805   3250 Template:Nornabr/switch
 12.51%  502.374  21764 Template:Sup
  6.26%  251.351      1 Page:An_Etymological_Dictionary_of_the_Norn_Language_in_Shetland_Part_I.pdf/130
  5.63%  225.997      1 Page:An_Etymological_Dictionary_of_the_Norn_Language_in_Shetland_Part_I.pdf/133
  5.04%  202.243      1 Page:An_Etymological_Dictionary_of_the_Norn_Language_in_Shetland_Part_I.pdf/139
  5.03%  202.060      1 Page:An_Etymological_Dictionary_of_the_Norn_Language_in_Shetland_Part_I.pdf/134
  4.87%  195.609      1 Page:An_Etymological_Dictionary_of_the_Norn_Language_in_Shetland_Part_I.pdf/129
  4.78%  192.097      1 Page:An_Etymological_Dictionary_of_the_Norn_Language_in_Shetland_Part_I.pdf/143
Do you need the complexity of Template:Nornabr. — billinghurst sDrewth 07:15, 20 June 2020 (UTC)
Ah, thanks for this. I have for now managed to optimise the switch in Template:Nornabr to roughly halve the post‐expand include size by re-ordering terms by how often they appear in the text. I will try to keep using it until it causes issue again, at which point I will consider other options. Thanks again.— 🐗 Griceylipper (✉️) 16:07, 20 June 2020 (UTC)
The option is to spawn out those that are really common to be their own abbreviations, or , or use LUA. #switch upon #switch is just expensive. — billinghurst sDrewth 02:10, 24 June 2020 (UTC)

lost my username[edit]

Please refresh my memory as to what username I use with Wikisource. —Preceding unsigned comment added by 2600:8803:7b80:a7b:1de9:d348:8d0a:fd5c (talk)

No way for us to know. Think of something that you edited, and look through the page's history for what you seek. — billinghurst sDrewth 02:07, 24 June 2020 (UTC)

Disambiguation of authors[edit]

I have questions regarding disambiguation of authors.

  1. There's no any author page with the name "James Gray" yet, though it's likely to have more than one author page with such name (see James Gray disam page at Wikipedia). So, if I am to create an author page for "James Gray (poet)", should I disam it for now?
  2. Help:Disambiguation says an author page is disambiguated by the author's years of life, and Wikipedia says "James Gray (poet)" was born circa 1770 (and died in 1830). If I have to disam his page, should I do it with "(c. 1770–1830)" (I mean with "c." or "circa" mentioned)?.

Thank you so much. --Asuracarya (talk) 04:28, 24 June 2020 (UTC)

Although there are quite a lot of people of this name at Wikipedia, it is quite likely that the one you want to add here will stay here alone for a long time. The Wikipedia disambiguation page was founded in 2013, while only now (7 years later) the first James Gray is going to be added to Wikisource. It may take 7 more years until another one will be added here. So it is not necessary to disambiguate in advance. Jan Kameníček (talk) 06:47, 24 June 2020 (UTC)
@Asuracarya: The community made the decision years ago to only disambiguate as necessary. So please just put it at the root, and we will disambiguate as/if required.

BUT if you were looking to do an external link, and concerned about the eventual outcome of page moves, then please create a redirect Author:James Gray (1770–1830) that points to the author page, and use that redirected link at the other place. Such redirects will get the maintenance required to point to the right place if we move things. AL THAT SAID the best way to maintain the links is through links at Wikidata as w:template:wikisource-author at English Wikipedia uses wikidata's wikilinks to generate links, and we already link that way from here. — billinghurst sDrewth 05:59, 25 June 2020 (UTC)

also noting Gray, James which should get linked to your pending author page. — billinghurst sDrewth 06:02, 25 June 2020 (UTC)
Seven years? strikes me as something worth preparing for Wikisource, as a second full translation of the Dhammapada is not a bad thing, nor a difficult one. James Gray (1849-1905) is the translator.--Prosfilaes (talk) 22:38, 26 June 2020 (UTC)

Footnotes across pages ending with hyphen[edit]

When a footnote extends over two pages and ends with a hyphen on the first page, it seems like the space inserted between the pages is not removed as usual. For example:

Page:Niti literature (Gray J, 1886).pdf/49 Page:Niti literature (Gray J, 1886).pdf/50 Transclusion result
<ref name = "abc">Ćâṇakya says: "To pay off debts, to quench a fire, and remove desire is good, for should they in-</ref> <ref follow = "abc">crease, they cannot be stopped."</ref> Ćâṇakya says: "To pay off debts, to quench a fire, and remove desire is good, for should they in- crease, they cannot be stopped."

Is there a way to fix this? Thank you very much. --Asuracarya (talk) 13:58, 27 June 2020 (UTC)

@Asuracarya: Try {{hws}}/{{hwe}}:
Page:Niti literature (Gray J, 1886).pdf/49 Page:Niti literature (Gray J, 1886).pdf/50 Transclusion result
<ref name = "abc">Ćâṇakya says: "To pay off debts, to quench a fire, and remove desire is good, for should they {{hws|in|increase}}</ref> <ref follow = "abc">{{hwe|crease|increase}}, they cannot be stopped."</ref> Ćâṇakya says: "To pay off debts, to quench a fire, and remove desire is good, for should they increase, they cannot be stopped."
It's a little tedious, but it gets the job done. --Xover (talk) 14:20, 27 June 2020 (UTC)

Style check requested...[edit]

User:ShakespeareFan00/CCE1 , Is this rendering properly? ShakespeareFan00 (talk) 11:19, 29 June 2020 (UTC)

Looks OK to me, though it could do with a padding-left:1em; or similar in column 2. The width:100% is also a bit unnecessary, IMO.
Are you sure plugging together several generic CSS classes from multiple stylesheets and then applying {{ts}}'s ad-hoc to cover whatever those classes don't is easier than just writing a single CSS sheet for the exact kind of table? The w:DRY principle is a good idea, but only if it actually makes things more robust. In this case, you could encode pretty much all the formatting, including the right-alignment of the second line in the right cell (which you have ignored), padding of col 2 and header and even the caption with CSS like this:

Example: CCE table CSS

._cce_table {
	width: 100%; /* but maybe margin: auto; */
	border-collapse: collapse;

._cce_table td {
	vertical-align: top;

._cce_table td:not(:first-of-type),
._cce_table th:not(:first-of-type) {
	border-left:1px solid black;
	padding-left: 1em;

._cce_table td:not(:last-of-type),
._cce_table th:not(:last-of-type) {
	padding-right: 1em;

._cce_table td:nth-child(2) p:last-of-type {
	margin: 0;

._cce_table th {
	font-weight: normal;
	font-variant: small-caps;
	border-top:1px solid black;
	border-bottom:1px solid black;
	padding: 0.5em 0;

._cce_table caption {
	font-size: 120%;
	text-align: center;
This means your table can be marked up with only a single class and no extra formatting:
<templatestyles src="THISCSS.css"/>
{| class="_cce_table"
|Boyd's Directory of Shamokin. 1891–1893||W. H. Boyd,

Reading, Pa.

and you also don't need to worry that someone might break the central CSS in some subtle way.
If you have "special" rows, you can add a class just for them to add/remove styles as needed, which is much easier to reason about when all the CSS is in one place and not sourced from snippets of several stylesheets. Inductiveloadtalk/contribs 12:23, 29 June 2020 (UTC)
If you want to implement a class for this work, feel free, This is a wiki after all :). And yes ideally I'd like to have the tables as simple as possible. ShakespeareFan00 (talk) 12:47, 29 June 2020 (UTC)

Template for linking to subpages in ToC[edit]

I know about {{TOC link}}, and it's great for linking page numbers in Tables of Contents -- but I'd also like to link the section name to the subpage (unconditionally). I can do it manually, but that means repeating the whole parent page name in each row, and repeating the subpage name twice (once as the name, once as the label). It feel like there should be a template to reduce that to something like {{TOC page link|This Is An Essay}} (which would become [[Book of Essays/This Is An Essay|This Is An Essay]]). Does this exist, or should I create it? JesseW (talk) 02:20, 2 July 2020 (UTC)

@JesseW: I've been thinking about that, but so far haven't come up with any good solution. The problem is that templates do not know anything about their calling context: there is no way to set the base page name once and have subsequent invocations reuse it, and there is no way to automatically guess it. Any solution to this would have to be based on JavaScript ala. the Easy LST gadget in your preferences (it transforms wikitext on load and on save, and JS can do much more than templates can). --Xover (talk) 09:35, 2 July 2020 (UTC)
We've been through the thinking before. We can set relative links in the ToC, though that leads to red links. The community discussed and pretty much resolved that it just easiest to set hard links. Hard links are not perfect either on the occasions that we need to works, though less problematic.

TOC links to the page namespace don't generate much. Sure we set them up to be done, though it is really of little value as there is zero progress in repopulating the works at Commons. — billinghurst sDrewth 12:50, 2 July 2020 (UTC)

The lack of an elegant technological solution offends my technology-optimist sensibilities. :-) --Xover (talk) 13:48, 2 July 2020 (UTC)

OK, so we have to specify the parent page ("Book of Essays" in my example). But we should still be able to avoid repeating the subpage. I've made such a template, {{subpage|A|B}} which becomes B; if such a template already exists (or exists in a cleaner way), please redirect (or just delete my version). JesseW (talk) 16:12, 2 July 2020 (UTC)

Hearing no objections or suggestions of an existing alternative, I'm going to start using it (and add documentation, etc.) JesseW (talk) 18:57, 4 July 2020 (UTC)

Question .. From a total Newbie .. who does not want to offend anyone![edit]

I have been studying the Treaty of Fort Laramie of 1868 - and the versions that exist online such as the National Archive ( ),,, etc., all read quite differently from this source. Is there a ‘version’ dispute, or “two schools” situation? I couldn’t find where to ask this question, so here goes .. JS. unsigned comment by Sgtpepper6344 (talk) 08:03, 5 July 2020‎.

The WS text hasn't been changed in content since 2006 (in the very early days of Wikisource), when it was imported from Wikipedia prior to deletion there: w:Wikipedia:Articles for deletion/Treaty of 1868. There is no stated source, and the text looks very rather like a dump from some third-party source.
A cursory comparison of the WS text and the scanned original's first page doesn't look radically different to me, except for some missing line breaks.
I have constructed and imported a scan (constructed from the images) at Index:Treaty of Fort Laramie - 1868.pdf, which can be used to provide a scan-backed work with greater integrity than an anonymous 14-year-old version. Inductiveloadtalk/contribs 12:00, 5 July 2020 (UTC)