T272319 closed as resolved.

exclusionRegex must be in editConfig

I was trying to use this out on a private MediaWiki wiki (one that requires a log in for any actions) and this fails with the following error:
::error::readapidenied: You need read permission to use this module.

I tried the same simple code, to login and read a page, on Wikipedia and it worked as expected so my code seems all good.

This seems similar to an issue I found with a python mediawiki library: barrust/mediawiki#79

For private wikis you must login before you can get the api info. I'm not as familiar with Node so my understanding of the details of your code are limited, but I think you are doing a similar request before logging in. So the solution may be similar if you want mwn to be able to be used with private wikis.

Thanks!

Hi, in certain scenarios I'd like to create an instance of Template directly. One such scenario are tests for wiki-specific functionality that relies on Template. As it is part of the documentation I expected this to work, however it currently is only exported as type so I've to rely on a workaround that requires me to setup a proper bot instance.

const wt = new bot.Wikitext("{{template}}");
wt.parseTemplates({});
const template = wt.templates[0];

To-do items are now being tracked on this project board.

Documentation, especially for the stuff in "Bulk-processing functions" section, is much needed.

Given that MW API treats protection levels absolutely and not relatively, this will have to implemented a bit smartly to be useful. See https://en.wikipedia.org/wiki/MediaWiki:Gadget-morebits.js for hints: functions this.protect and fnProcessProtect.

Wikitext of [[:en:Greg Egan]]:

{{Infobox writer <!-- for more information see [[:Template:Infobox writer/doc]] -->
| name        = Greg Egan
...

Mwn will list the following template names on the above page: Short description, Infobox writer <!-- for more information see [[:Template:Infobox writer/doc]] -->, Cite journal, Cite web, ...

This package has been super useful to me, thank you for making it!
I recently stumbled upon an unusual use case, and I'm not sure how to handle it:

• I am logged in to, say enwiki, using BotPassword.
• I do not have a local account on, say frwiki.
• If I attempt to login to frwiki with BotPassword, my login will be rejected (Incorrect username and password).
• If I try and fetch a CentralAuthToken from enwiki by hand, in order to use it in my request to frwiki (see CentralAuth/API), I get the following error:

code: 'badsession',
info: 'Can only obtain a centralauthtoken when using CentralAuth sessions.'

What is the proper way to interact with other wikis where I may not have a local account?

Copying off from the list at https://www.mediawiki.org/wiki/API:Client_code/Gold_standard

Easy to install

• Installation instructions are correct and easy to find
• Library is packaged for installation through appropriate package library (PyPI, CPAN, npm, Maven, rubygems, etc.)
• Platinum standard: library is packaged for and made available through Linux distributions

Easy to understand

• Well designed: makes all intended API calls available with the intended level of abstraction with no redundancies
• Platinum standard: makes the Wikidata API available

Well documented

• Code is commented and readable
• Documentation is comprehensive, accurate, and easy to find
• Deprecated functions are clearly marked as such
• Platinum standard: Documentation is understandable by a novice programmer
• Code uses idioms appropriate to the language the library is written in

Easy to use

• Has functioning, simple, and well-written code samples for common tasks
• Demonstrates queries
• Demonstrates edits

Handles API complications or idiosyncrasies so the user doesn't have to:

• Login/logout
• Cookies
• Tokens
• Query continuations using the new "continue" and not "query-continue"
• (?) Requests via https, including certificate validation
• Courteous API usage is promoted through code samples and smart defaults
• gzip compression is used by default
• Examples show how to create and use a meaningful user-agent header (as in meta.wikimedia.org/wiki/User-agent_policy)
• Platinum standard: generates a unique user-agent string given name/email address/repository location
• Efficient usage of API calls
• Can be used with the most recent stable version of the language it is written in (e.g. Python 3 compatible)

Easy to debug

• Contains unit tests for the longest and most frequently modified functions in the library
• Platinum standard: Unit tests for many code paths exist and are maintained
• Terrible hacks/instances of extreme cleverness are clearly marked as such in comments
• Documentation links to the relevant section/subpage of the API documentation

Easy to improve

Many of these aren't really applicable -- I have marked such items as checked because there is nothing to be done about them

• Library maintainers are responsive and courteous, and foster a thoughtful and inclusive community of developers and users
• Platinum standard: Project sets clear expectations for conduct[1][2] for spaces where project-related interactions occur (mailing list, IRC, repository, issue tracker). It should:
• State desired attitudes and behaviors
• Provide examples of unwelcome and harassing behavior
• Specify how these expectations will be enforced
• Pull requests are either accepted or rejected with reason within 3 weeks (Platinum standard: 3 business days)
• Issues/bugs are responded to in some manner within 3 weeks (Platinum standard: 3 business days) (but not necessarily fixed)
• The library is updated and a new version is released within 3 weeks (Platinum standard: 3 business days) when breaking changes are made to the API
• Platinum standard: library maintainers contact MediaWiki API maintainers with feedback on the API's design and function
• Library specifies the license it is released under

While the comment states that ns is part of the return object for MwnFile,usages, the typing doesn't specify it. As the API returns the value I'd guess the return type is wrong.

Similar to mw.Title. Useful for normalising page titles, getting talk page name from the page name, etc.

Will need to think about how to best handle the fetching of data about namespace names.

To upload files from URL. See https://www.mediawiki.org/wiki/API:Upload#Example_2:_Upload_file_from_URL

Hello, I'd like to do a cargo query to a fandom.
Here is the query : https://lol.fandom.com/api.php?action=cargoquery&tables=Tournaments&fields=Tournaments.Name,Tournaments.OverviewPage

I have no clue on how to do it with any MediaWiki js api
As this is the latest updated one I'm seeking for help here

Nothing found on internet

Do someone know how to do it ?

Can one also use events? So, when someone edits, protects, deletes a page, blocks, unblocks users, grants/revokes permissions, etc?

In our wiki we have some categories specifically to group templates. In certain scenarios however this seems to cause issues with the bot. Ie new bot.Category("Template:Infobox") throws Error: not a category page instead of referencing Category:Template:Infobox. This can also be observed with MwnTitle: new bot.Title('Template:Infobox', Namespace.CATEGORY); results in { namespace: 10, title: 'Infobox', fragment: null }. Using an invalid prefix however works as intended: new bot.Title('Invalid:Infobox', Namespace.CATEGORY); results in { namespace: 14, title: 'Infobox', fragment: null }. Prefixing the title correctly with Category: also works: new bot.Title('Category:Template:Infobox', Namespace.CATEGORY); results in { namespace: 14, title: 'Template:Infobox', fragment: null }.

One use case where this issue becomes apparent is when trying to map included categories:

    const article = new bot.Page(PAGE_IN_TEMPLATE_CATEGORY, Namespace.TEMPLATE);
    const categories = (await article.categories()).map(({ category }) => new bot.Category(category));

The history and category methods seem particularly interesting.

See table in :mw:Manual:Creating_a_bot

The maxlagPause option is missing in the MwnOptions object.

Maybe you meant the retryPause option?

Hi,

There seems to be an issue with the mwn class consisting of some methods that return void or Promise#void when they should in fact return data from the API. An example of this is mwn#getSiteInfo, which returns void since it's not returning the result of Title.processNamepsaceData (which is also returning void).

See mwn#getSiteInfo's code:

getSiteInfo(): Promise<void> {
    return this.request({
        action: 'query',
        meta: 'siteinfo',
        siprop: 'general|namespaces|namespacealiases',
    }).then((result) => {
        // Title.processNamespaceData should have a return statement
        // should also be a return statement here
        this.title.processNamespaceData(result);
    });
}

You could also just use .then(this.title.processNamespaceData) for more concise syntax.

Types may also need to be adjusted to suit any changes. Please let me if I am misunderstanding the usage of this function.

Thanks for the great library. I just noticed that there's a typo in bot.ts#L750: it's "apihighlimits", not "apihighlimit", without "s". I was wondering why bot.hasApiHighLimit always returns false even though my bot has a bot flag, but this seems to be the cause.

Executing backlinks on an missing article throws MwnErrorMissingPage, however the API properly returns the links to the missing article. This would be useful in certain scenarios, ie pages that were moved without a redirect where a bot now should adjust links.

We need to be clear about what versions of MediaWiki and Node.js are supported.

Now new Mwn().Wikitext.parseSections('') (empty text) returns one element with empty content string:

[
    {
        level: 1,
        header: null,
        index: 0,
        content: ''
    }
]

I think parseSections('') must returns the empty array or null.

Hi,

in my scenario there are also automated bot task where I want to write output to a file (or something else) instead of using stdout. So it'd be nice to have the possibility to overwrite the internal log method to integrate especially the API warnings to my normal logging system.

Hi, I recently got this:

Error [ERR_PACKAGE_PATH_NOT_EXPORTED]: Package subpath './lib/defaults' is not defined by "exports" in /home/Scripts/IPChecker/node_modules/axios/package.json

New user, bit confused. Is there a version 2 release?
In the getting started guide it says:

Prior to mwn v2.0.0, import was via const { mwn } = require('mwn'). Prior to v0.8.0, it was via const mwn = require('mwn');

But the latest release as of writing is 1.11.5. Is this a typo? I'm confused.

The request library used for making HTTP requests is deprecated. It needs to be replaced with another libarary, like axios.

None of the links for this package's documentation work on the NPM page. Is there any link for documentation that works?

I have a simple article that redirects to a missing article. When creating the page instance and immediately calling getRedirectTarget or isRedirect the call throws a MwnErrorMissingPage error because the api (obviously) returns that the target doesn't exist. However when first calling text on the page instance, getRedirectTarget uses the text of the article to determine the (intended) redirect target and correctly returns it. Imho the (externally observed) behaviour shouldn't change based on other calls I might've done before, ideally returning the target either way.

While checking the function I've seen that there's also a regex in place, however that didn't work at all for me and it also doesn't consider localizations (eg in a german wiki #weiterleitung and #redirect are valid).

Hey, I've been updating some dependencies across our projects and noticed that mwn still uses version ^0.25.0 of axios. The latest version for axios is 1.6.2. Several vulnerabilties have been fixed between 0.25.0 and 1.6.2, which makes it worthwhile to upgrade.

The latest version of axios is currently incompatible with mwn as-is, but may be fixed with some small code changes.

If a page isn't linked anywhere calling backlinks throws an error: TypeError: Cannot read properties of undefined (reading 'map'). Seems like page.linkshere should fall to an empty array in that case.

Processing templates in wikitext is a frequent bot requirement. While API:Parse lets us know the templates present in wikitext, it recursively includes the templates within the templates, and doesn't tell us what parameters were used.

Port the stuff from https://en.wikipedia.org/wiki/MediaWiki:Gadget-libExtraUtil.js written by @evad37 and myself.

Looks like a very polished tool, both for docs, code comments, and typed source code--nice!

Just wanted to suggest that for servers with CORS enabled, a browser client should be able to do pretty much anything that Node can do in the context of a Bot, so wanted to raise the possibility of having Node and browser entrance files, where the Node one would pass along node-fetch and the browser one, pass along window.fetch, etc., and the API would make use of them so one could use your code solely in browser-side code too.

For my purposes at least, I think it should be easier to just write client-side code (which I need for a GUI anyways).

Might not be something on your priority list, but wanted to express to give it some thoughts, if nothing else as an eventual goal perhaps. Thanks and best wishes!

Checklist of before-hands:

• Use Docker for automated testing of write actions in CI
• Using above, check compatibility going back to MW 1.34 LTS version
• Expand test coverage to 80%+
• Make error handling more consistent
• Better documentation, especially for nested classes
• Rename all classes to follow PascalCase convention

Currently, mwn does not attempt to intercept expired sessions to perform re-logins if the session expired (usually after being logged in for extended periods). Unfortunately, this bleeds into request functions such as massQuery, and ends up causing a chain of errors down the line due to the unexpected input (a login error instead of the expected response).

For reference, here is the data returned by one of the massQuery responses:

{
    "code": "mwn_failedlogin",
    "info": "Failed: Unable to continue login. Your session most likely timed out.",
    "response": {
        "login": {
            "result": "Failed",
            "reason": "Unable to continue login. Your session most likely timed out."
        }
    }
}

Subsequent errors after this one return the following:

{
    "code": "mwn_failedlogin",
    "info": "Login failed",
    "response": {
        "login": {
            "result": "WrongToken"
        }
    }
}

It would be great to have mwn automatically check and determine if the session has expired and seamlessly re-login with a new login token. This removes the need to pile on additional code on every single request in order to trigger a re-login (and re-try of the request).