The AI Engine

how to use plugins.
example 1
keyphrase = is (p:emotions2) a emotion
response=yes

example 2
keyphrase = i am (p:emotions2) right now.
response = i sometimes feel that way.

play around with that for now.
i will answer the other questions ounce you get the hang of that.

When it comes to names, you can ask the user what their name is, and have a response of xnomatch so it catches any response. Add the aiscript rem 'name' as only (key1), then use (mem-name) any time you want to use their name.

Also, I'm finding that the lowercase seeks work fine with uppercase and mixed case responses - unless you're using raw aiscript, the site turns all messages to lowercase when it does its spellcheck etc.. what was failing to be matched, and by what? When using debug mode, did it say the input was still uppercase after processing?

yes and no's tend to need to be raw to work. not always the case though. there was something else like this too but idk

Thank you for responses, I will play around with what was mentioned. I'm still having trouble wrapping my mind around (Key)'s but I don't have any specific question except the general concept of not understanding. But I'll try and reread the book of AI again. Also, the response that wasn't working was "hey" it wouldn't understand "Hey" so I had to make a word list. I'm finding myself making word lists with every possible variation of the word and I'm not using raw. But again thank you for responding, I really appreciate it.

Hey is odd, I believe. I think all greetings get pulled into xhello, and triggers once per person before triggering xgreeting instead.

I'm guessing no one knows anything about my plug-in failure ?

to recap im talking about how in convo's, used plugins don't always seem to work. so the bot just ends up saying a incomplete sentence.

I could just replace most of them with (asdas|asdas) but it would kinda defeat the point.

iv also noticed the rem (mem-anything) as only "w/e" is buggy.. its made a few new mems like as only "yes" and it makes a new mem called yes and remembers it as the next as only requested.. might have to go back and change it.. just a heads up.

But I seriously need to know about that plug in bug..

i would not use slash with memory

yeahhh w/e just means "what ever" lol but I'm seeing it allot more, were rem (mem-bob) as only "1" some how becomes (mem-1) as only 2 or some random crap.. its showing up allot now during tests, ill probably switch to "1" as only "bob".. but fuck.. honestly like 80-90% of my kp have mems attached like that.. I could just keep replacing them as I notice it but idk.. I'm tired lol

it definitely seems like its confusing what its supposed to remember with what its next memory is supposed to be.. but on paper it all looks sound, and sometimes what its replaced with seems random.. it shouldn't have any trouble.. so I have to assume its a bug..

it could also be conflicting with something some how, like when i had a xnomatch and a seek looking for different versions of the same thing broke them both in the intro. kinda wish i woulda wrote that down, cause it feels allot like that but instead of replacing the mem with a kp its replacing it with another mem.. maybe..? in debug.. it looks like its creating a post key seek for the mem, or i could just be miss understanding what I'm seeing.. but like i said idk.

ill try some more in debug when i get the chance..

getting a little disheartened between this and that plug in problem.. guessing no one else is having that issue since no ones said anything about it. maybe I'm the only one using plug-ins in responses.. idk

will admit though debug is pretty kool, it can also show break downs and possible variations for statements.
I could learn a thing or two about wild cards from it. there was also a $k at the end of them, does that mean anything..to the ai engine?

i think it does not matter if (mem-bob) as only "2" or any other number.
so long as the keyphrase triggers the response.

what keyphrase for what sentence are you trying to make with a plugin?