pfaffle / rawrbot Goto Github PK

If a user sends a private message to the bot, it should default to returning 'Learning' data to the user, unless the bot's prefix is detected.

For example, desired behavior would be (In a PM window with the bot):

<user> foo
<bot> foo is bar.
<user> !ldap bob
<bot> Performing LDAP search on bob...
...

Add commands to make the bot join or part a particular channel, e.g.
!join #channel [channelkey]
!part #channel

stenlai's infoobot entry is getting out of hand. I suggest adding a command that allows for archiving to Chronicle.

Would be nice to have some unit tests! I started looking at https://rubygems.org/gems/cinch-test/versions/0.1.1 a while ago, which might be useful for this.

The RT plugin stopped working a while ago on OIT's RT instance. This is a reminder to me to fix it.

The instructions fail to mention that rvm needs to be installed and configured to start running the bot.

Right now, the prefix '!' for commands is hardcoded in a number of different places (plugins). These instances need to be removed and made to use the prefix defined in the main Rawrbot config file.

Right now the bot just does whatever it is told, regardless of who is asking, as long as the command is valid. I would like to add some access controls for some of the functions (such as join/part, send message, etc.) so that only bot "admins" can tell the bot to use them.

The config file defines an array called "admins" which contains nicks of the people who should have access to more sensitive functions, but none of the plugins currently use that array to check if a user is authorized to perform those functions. This needs to be implemented.

When fetching the title of a webpage, HttpTitle doesn't follow redirects (e.g. from http -> https URLs). It should.

When adding onto a learning entry something that begins with a pipe '|' character, rawrbot should not prepend the entry with 'or' like it does normally. That way, you can easily continue to build off an existing entry that is purely made up of pipe-separated values. Example of desired behavior:

<pfaffle> rawrbot: literal foo
<rawrbot> foo =is= thing1|thing2
<pfaffle> rawrbot: foo is |thing3
<rawrbot> OK, pfaffle.
<pfaffle> rawrbot: literal foo
<rawrbot> foo =is= thing1|thing2|thing3

Current behavior after adding 'thing3' to the entry is:

<rawrbot> foo =is= thing1|thing2 or |thing3

The Social plugin is currently pretty minimal and could be expanded. Some potential features to add include:

• Ability to recognize more greetings/goodbyes.
• Additional variety in responses.
• Ability to recognize other kinds of common small talk questions and respond to them.

This might require some additional intelligence in this plugin and/or the Learning plugin in order to better recognize which one of them should be triggered. It's also possible that both plugins might need to be merged.

With the plugins set to [Social,Messenger,Karma,Learning,GoogleRSS], which should mean that no LDAP configs are needed, I get the following error:

~/Rawrbot/lib/ldap_helper.rb:18:in `read': No such file or directory - config/ldap.yml (Errno::ENOENT)
        from ~/Rawrbot/lib/ldap_helper.rb:18:in `initialize'
        from ~/Rawrbot/plugins/USSCATldap_plugin.rb:23:in `new'
        from ~/Rawrbot/plugins/USSCATldap_plugin.rb:23:in `<class:USSCATldap>'
        from ~/Rawrbot/plugins/USSCATldap_plugin.rb:15:in `<top (required)>'
        from ~/.rvm/rubies/ruby-1.9.3-p484/lib/ruby/site_ruby/1.9.1/rubygems/core_ext/kernel_require.rb:55:in `requ
ire'
        from ~/.rvm/rubies/ruby-1.9.3-p484/lib/ruby/site_ruby/1.9.1/rubygems/core_ext/kernel_require.rb:55:in `requ
ire'
        from ./rawrbot.rb:22:in `block in <main>'
        from ./rawrbot.rb:21:in `each'
        from ./rawrbot.rb:21:in `<main>'

<me> Rawrbottest
<Rawrbot> bugger all, I dunno

It should only trigger if there is a space between the bot's name and the thing being looked up

The Twitter plugin is broken right now since it doesn't implement OAUTH and the Twitter v1.1 API. This needs to be updated to work.

Ruby 1.9 is getting old. Test Rawrbot in Ruby 2.1 and update any code that doesn't work with it.

When the bot tries to identify if it is being addressed by name, it should not be case-sensitive.