User:Lowercase sigmabot III/Archive HowTo

Since November 2013, Lowercase sigmabot III (talk · contribs) has taken over all tasks that used to be performed by MiszaBot. The configuration syntax is unchanged, which is why you will see the name "MiszaBot" included in configuration instructions below. According to the source code, Lowercase sigmabot III only uses {{User:MiszaBot/config}} as the config. You cannot substitute "Lowercase sigmabot III" for "MiszaBot" in the configuration.

For similar bots, see Category:Wikipedia archive bots.

Before setting up automatic archiving on a high-traffic article's talk page, please establish a consensus that archiving is necessary there.

If you are unsure what code to use, a suggestion: go to #Example 2: Incremental archives and have a look at its copy paste section.

Setting up archiving

To quickly setup archiving, place {{subst:Setup auto archiving}} above the first heading. You'll get incremental archives, much like in Example 2 below.

Put the following template at the top of the page that you want automatically archived:

{{User:MiszaBot/config
| algo    = old(...)
| archive = Name of archive page(s): To generate multiple pages over time, use variables, described below.
[other parameters - ***read below***]
}}

Well, maybe not exactly that (read about required options below), but that's how it's done. By the way, the template has no content and is template protected, so don't worry about visual appearance. :)

The bot expects to see exactly one parameter per line. The closing }} must be on its own line. There must be no leading spaces before the pipes, and no blank lines. If you drop out the newlines used in the above example, e.g. by wrapping the template invocation onto a single line, the bot may not do what you expect.

Also, the template must be located:

1. before the first ==second level header== on the page
2. on the actual page and not transcluded from a subpage, such as a page header, etc.

Quick examples

To get the hang of it – configuration may not be intuitive, but it's flexible and gives potentially vast possibilities.

Please change "User talk:Example" to the exact name of your own page!

If it's still not working, read about the "minthreadsleft" and "minthreadstoarchive" parameters below.

Example 1: Static archive

{{User:MiszaBot/config
| algo    = old(7d)
| archive = User talk:Example/Archive
}}

(Note: this is an example and this exact code won't work on your page.)

This setup archives threads from User talk:Example to User talk:Example/Archive after they are 7 days old (counting from newest timestamp). The archive will freeze when it becomes 2000 KB large. It is probably not a very useful config (unless for sending threads to a periodically cleaned trash bin or to decide yourself when to go to the next archive page), but it is easiest, so it is mentioned first.

Example 2: Incremental archives

{{User:MiszaBot/config
| algo                = old(30d)
| archive             = User talk:Example/Archive %(counter)d
| counter             = 1
| maxarchivesize      = 150K
| archiveheader       = {{Archive}}
| minthreadstoarchive = 1
| minthreadsleft      = 4
}}

(Note: the "Example" code above won't work as is; please substitute in your userid, or just copy-paste the code below!)

copy and paste for easy use

{{User:MiszaBot/config
| algo                = old(30d)
| archive             = {{SUBST:#titleparts:{{SUBST:FULLPAGENAME}}}}/Archive %(counter)d
| counter             = 1
| maxarchivesize      = 150K
| archiveheader       = {{Archive}}
| minthreadstoarchive = 1
| minthreadsleft      = 4
}}

This tells the bot to archive threads over thirty days old (leaving the four most recent) from User talk:Example to User talk:Example/Archive 1 (more about variables below) until it fills up to 150 kilobytes, whereupon the bot will move to 2 (updating the counter when saving page). Remember to specify the maximum size of an archive, or it will behave pretty much like in the first example. In addition, each archive page is given a {{Aan}} banner, which makes it easy to move between the different archive pages.

Example 3: Date-based archives

{{User:MiszaBot/config
| algo          = old(5d)
| archive       = User talk:Example/Archives/%(year)d/%(monthname)s
| archiveheader = {{Monthly archive list}}
}}

(Note: please substitute in your userid, or just copy-paste the code below!)

copy and paste for easy use

{{User:MiszaBot/config
| algo          = old(5d)
| archive       = {{SUBST:#titleparts:{{SUBST:FULLPAGENAME}}}}/Archives/%(year)d/%(monthname)s
| archiveheader = {{Monthly archive list}}
}}

In this configuration, threads older than 5 days will land in archives depending on their date (that is newest timestamp). You can read about other variables below.

Parameters explained

Note: For those parameters with defaults they will work regardless of whether they are listed on the talk page. But it is better to list all those parameters (with values of your choice) to avoid confusion. See this discussion.

{{User:MiszaBot/config
| archive             = 
| algo                = 
| counter             = 
| maxarchivesize      = 
| minthreadsleft      = 
| minthreadstoarchive = 
| archiveheader       = 
| key                 = 
}}

Parameter	Default	Description
archive	N/A; required	Name of the page to which archived threads will be put. This parameter supports the use of variables, which can be used to create dynamically named archives (such as using date ranges). Note that magic words and templates do not work with this parameter, and unless a key is supplied (see below), the target page must be a subpage of the current page. Otherwise Category:Pages where archive parameter is not a subpage is added.
algo	old(24h)	Minimum amount of time that must pass after the last reply before the bot can archive the thread. It must be in the form: old(nT) where n is a number and T specifies hours or days. You can leave out T to specify seconds (even there's essentially no use case for that). Examples: 600, 72h, 5d would specify either 600 seconds, 72 hours or 5 days respectively. To specify very long timespans, use days. For example; algo=old(1100d) for close to three years. Complex formulas (including regex matching) may be available in the future.
counter	1	The current number of the last archive. If the %(counter)d variable is not used, it is ignored. Lowercase sigmabot III will update this parameter as necessary.
maxarchivesize	1954K	The maximum archive size in either raw text size (eg 256M for megabytes, 256K for kilobytes, or 256B for bytes) or threads (eg 10T) before incrementing the counter. A bare integer will be interpreted as bytes. Ignored if counter is not used.
minthreadsleft	5	The minimum number of threads that should be left on a page (to prevent pages from getting completely harvested).
minthreadstoarchive	2	The minimum number of threads to archive at one time, which is used to lower edit frequency. lowercase sigmabot III will not archive threads when fewer than the value of this parameter would be archived.
archiveheader	{{Archive}}	Content that will be put on new archive pages as the header. This parameter supports the use of variables.
key		A top-secret key that (if valid) allows archives to not be subpages of the page being archived. To obtain such a key, request it from this editor, and you better have a very good justification for doing so.

Variables

Variable parameter	Expands to...
%(counter)d	the current value of the counter
%(year)d	year of the thread being archived
%(month)d	month (as a number 1-12) of the thread being archived
%(monthname)s	English name of the month above
%(monthnameshort)s	first three letters of the name above

Those python-savvy people will immediately recognize these variables are filled at runtime with the % operator and should know how to adjust the format. A few points of interest to others:

• make sure you don't mess up the last letter after the brackets (it's "d" for integers and "s" for strings)
• integer variables may be left-padded with zeros:

  %(counter)03d evaluates to 013 if the counter's value equals 13,

  %(month)02d becomes 05 for May etc.

After you have set up archiving

The bot runs once a day at a preset hour, or twice a day for WP:AN and WP:AN/I. Simply wait for the next cycle and you should see the bot's entries in the history list of your article providing there was anything to archive, and the history list entry should provide a link to the archive page you specified.

Various templates will produce an automatic list of archive subpages. Typically, such a template will be added to the top-level talk page to make archives easier to find. Simply add the template name in {{curly-braces}} where you want it to appear. Popular templates used for this purpose include:

• {{Archives}}
• {{Monthly archive list}}. If you use #Example 3: Date-based archives, use this with root = User talk:Example/Archives.
• {{Talk header}} (this template has other functions, but can include a list of archives and a search box)

A related but distinct function is templates that appear in the archived talk pages themselves. For instance:

• {{Archive}}

Delaying or preventing archiving of particular threads

Archiving can be delayed for a particular thread by substituting the template {{DNAU}} into the thread. Use {{subst:DNAU}} to retain a thread for a long time, or {{subst:DNAU|<integer>}} to retain a thread for <integer> days. Please see the template documentation for details about its use and function.

Missing signatures

One potential problem is that lowercase sigmabot III does not archive sections that have no signatures. However editors sometimes forget to sign a comment so it may ignore old sections while archiving more recent ones.

One option is to manually archive sections that are unintentionally unsigned. The other option, which allows lowercase sigmabot III to archive an unsigned section is to add a signature of the editor to the last comment in the section.

You can add missing signatures to the talk page by using the templates {{subst:unsigned}}, {{subst:unsigned2}}, {{subst:unsigned IP}}, or {{subst:xsign}}. Only the last comment added to a section needs to have a signature for lowercase sigmabot III to archive the section. This is now automatically added by User:SineBot.

If the talk page history is large, it may prove time consuming to find when a specific comment was added. WikiBlame will reduce the time it takes to find who added an unsigned comment.

Template output

When parameter |key= is provided, the only output of the template is categorization into Category:Pages archived using a key. Otherwise, the template outputs a single self closed tag <nowiki />. This is a trick, which makes it easier to deal with potential vertical white-space issues.

Causes of no archiving

Some of the possible reasons for a page not being archived:

• |archive= does not specify a subpage of the page to be archived. This can for example happen after a move where the parameter needs updating.
• Archive bot has not run. Archiving usually runs once a day. See the latest activity at Special:Contributions/Lowercase sigmabot III.
• One of the sections to be archived has an external link matching MediaWiki:Spam-blacklist or commons:MediaWiki:Spam-blacklist. Nothing will be archived. Manual archiving will also fail since blacklisted links cannot be saved. They can be deactivated with <nowiki>...</nowiki>.
• |minthreadstoarchive= is larger than the current number of sections ready for archiving.
• |minthreadsleft= is too large to currently allow any archiving.
• |algo=old(...) is currently too long for any archiving.
• The archiving instructions are not placed somewhere before the first section heading.
• The archiving instructions are placed within another template's code.
• The parameters are not arranged with exactly one per line.
• If some sections are archived but not others then they may have no signature with a valid time stamp. Please see #Missing signatures above.
• If your talk page's title contains certain punctuation characters ("&'). Please see Help:Archiving a talk page#FULLPAGENAME for an explanation and solution.
• If the topic has no subject line, such as welcome messages.