Help System

> help theatre
Help for The theatre (aliases: the hub, curtains)

The theatre is at the centre of the city, both literally and figuratively ...
(A lot more text about it follows ...)

  theatre/dramatis personae
> help evennia
No help found

There is no help topic matching 'evennia'.
... But matches where found within the help texts of the suggestions below.

  grapevine2chan, about, irc2chan

Evennia has an extensive help system covering both command-help and regular free-form help documentation. It supports subtopics and if failing to find a match it will provide suggestsions, first from alternative topics and then by finding mentions of the search term in help entries.

The help system is accessed in-game by use of the help command:

help <topic>

Sub-topics are accessed as help <topic>/<subtopic>/....

Working with three types of help entries

There are three ways to generate help entries:

  • In the database

  • As Python modules

  • From Command doc strings

Database-stored help entries

Creating a new help entry from in-game is done with

sethelp <topic>[;aliases] [,category] [,lockstring] = <text>

For example

sethelp The Gods;pantheon, Lore = In the beginning all was dark ...

This will create a new help entry in the database. Use the /edit switch to open the EvEditor for more convenient in-game writing (but note that devs can also create help entries outside the game using their regular code editor, see below).

The HelpEntry stores database help. It is not a Typeclassed entity and can’t be extended using the typeclass mechanism.

Here’s how to create a database-help entry in code:

from evennia import create_help_entry
entry = create_help_entry("emote",
                "Emoting is important because ...",
                category="Roleplaying", locks="view:all()")

File-stored help entries

New in version 1.0.

File-help entries are created by the game development team outside of the game. The help entries are defined in normal Python modules (.py file ending) containing a dict to represent each entry. They require a server reload before any changes apply.

  • Evennia will look through all modules given by settings.FILE_HELP_ENTRY_MODULES. This should be a list of python-paths for Evennia to import.

  • If this module contains a top-level variable HELP_ENTRY_DICTS, this will be imported and must be a list of help-entry dicts.

  • If no HELP_ENTRY_DICTS list is found, every top-level variable in the module that is a dict will be read as a help entry. The variable-names will be ignored in this case.

If you add multiple modules to be read, same-keyed help entries added later in the list will override coming before.

Each entry dict must define keys to match that needed by all help entries. Here’s an example of a help module:

# in a module pointed to by settings.FILE_HELP_ENTRY_MODULES

    "key": "The Gods",   # case-insensitive, can be searched by 'gods' too
    "aliases": ['pantheon', 'religion']
    "category": "Lore",
    "locks": "read:all()",  # optional
    "text": '''
        The gods formed the world ...

        # Subtopics

        ## Pantheon

        The pantheon consists of 40 gods that ...

        ### God of love

        The most prominent god is ...

        ### God of war

        Also known as 'the angry god', this god is known to ...

    "key": "The mortals",


The help entry text will be dedented and will retain paragraphs. You should try to keep your strings a reasonable width (it will look better). Just reload the server and the file-based help entries will be available to view.

Command-help entries

The __docstring__ of Command classes are automatically extracted into a help entry. You set help_category directly on the class.

from evennia import Command

class MyCommand(Command): 
    This command is great! 

      mycommand [argument]

    When this command is called, great things happen. If you 
    pass an argument, even GREATER things HAPPEN!


    key = "mycommand"

    locks: "cmd:all();read:all()"   # default 
    help_category = "General"       # default
    auto_help = True                # default 

    # ...

When you update your code, the command’s help will follow. The idea is that the command docs are easier to maintain and keep up-to-date if the developer can change them at the same time as they do the code.

Locking help entries

The default help command gather all available commands and help entries together so they can be searched or listed. By setting locks on the command/help entry one can limit who can read help about it.

  • Commands failing the normal cmd-lock will be removed before even getting to the help command. In this case the other two lock types below are ignored.

  • The view access type determines if the command/help entry should be visible in the main help index. If not given, it is assumed everyone can view.

  • The read access type determines if the command/help entry can be actually read. If a read lock is given and view is not, the read-lock is assumed to apply to view-access as well (so if you can’t read the help entry it will also not show up in the index). If read-lock is not given, it’s assume everyone can read the help entry.

For Commands you set the help-related locks the same way you would any lock:

class MyCommand(Command):
    <docstring for command>
    key = "mycommand"
    # everyone can use the command, builders can view it in the help index
    # but only devs can actually read the help (a weird setup for sure!)
    locks = "cmd:all();view:perm(Builders);read:perm(Developers)

Db-help entries and File-Help entries work the same way (except the cmd-type lock is not used. A file-help example:

help_entry = {
    # ...
    locks = "read:perm(Developer)",
    # ...

Changed in version 1.0: Changed the old ‘view’ lock to control the help-index inclusion and added the new ‘read’ lock-type to control access to the entry itself.

Customizing the look of the help system

This is done almost exclusively by overriding the help command

Since the available commands may vary from moment to moment, help is responsible for collating the three sources of help-entries (commands/db/file) together and search through them on the fly. It also does all the formatting of the output.

To make it easier to tweak the look, the parts of the code that changes the visual presentation and entity searching has been broken out into separate methods on the command class. Override these in your version of help to change the display or tweak as you please. See the api link above for details.


New in version 1.0.

Rather than making a very long help entry, the text may also be broken up into subtopics. A list of the next level of subtopics are shown below the main help text and allows the user to read more about some particular detail that wouldn’t fit in the main text.

Subtopics use a markup slightly similar to markdown headings. The top level heading must be named # subtopics (non case-sensitive) and the following headers must be sub-headings to this (so ## subtopic name etc). All headings are non-case sensitive (the help command will format them). The topics can be nested at most to a depth of 5 (which is probably too many levels already). The parser uses fuzzy matching to find the subtopic, so one does not have to type it all out exactly.

Below is an example of a text with sub topics.

The theatre is the heart of the city, here you can find ...
(This is the main help text, what you get with `help theatre`)

# subtopics

## lore

The theatre holds many mysterious things...
(`help theatre/lore`)

### the grand opening

The grand opening is the name for a mysterious event where ghosts appeared ...
(`this is a subsub-topic to lore, accessible as `help theatre/lore/grand` or
any other partial match).

### the Phantom

Deep under the theatre, rumors has it a monster hides ...
(another subsubtopic, accessible as `help theatre/lore/phantom`)

## layout

The theatre is a two-story building situated at ...
(`help theatre/layout`)

## dramatis personae

There are many interesting people prowling the halls of the theatre ...
(`help theatre/dramatis` or `help theathre/drama` or `help theatre/personae` would work)

### Primadonna Ada

Everyone knows the primadonna! She is ...
(A subtopic under dramatis personae, accessible as `help theatre/drama/ada` etc)

### The gatekeeper

He always keeps an eye on the door and ...
(`help theatre/drama/gate`)

Technical notes

Help-entry clashes

Should you have clashing help-entries (of the same name) between the three types of available entries, the priority is

Command-auto-help > Db-help > File-help

The sethelp command (which only deals with creating db-based help entries) will warn you if a new help entry might shadow/be shadowed by a same/similar-named command or file-based help entry.

The Help Entry container

All help entries (no matter the source) are parsed into an object with the following properties:

  • key - This is the main topic-name. For Commands, this is literally the command’s key.

  • aliases - Alternate names for the help entry. This can be useful if the main name is hard to remember.

  • help_category - The general grouping of the entry. This is optional. If not given it will use the default category given by settings.COMMAND_DEFAULT_HELP_CATEGORY for Commands and settings.DEFAULT_HELP_CATEGORY for file+db help entries.

  • locks - Lock string (for commands) or LockHandler (all help entries). This defines who may read this entry. See the next section.

  • tags - This is not used by default, but could be used to further organize help entries.

  • text - The actual help entry text. This will be dedented and stripped of extra space at beginning and end.

Help pagination

A text that scrolls off the screen will automatically be paginated by the EvMore pager (you can control this with settings.HELP_MORE_ENABLED=False). If you use EvMore and want to control exactly where the pager should break the page, mark the break with the control character \f.

Search engine

Since it needs to search so different types of data, the help system has to collect all possibilities in memory before searching through the entire set. It uses the Lunr search engine to search through the main bulk of help entries. Lunr is a mature engine used for web-pages and produces much more sensible results than previous solutions.

Once the main entry has been found, subtopics are then searched with simple ==, startswith and in matching (there are so relatively few of them at that point).

Changed in version 1.0: Replaced the old bag-of-words algorithm with lunr package.