AutoChat (In Dev)

[xxpoopyxx]

I own the server so I can put whatever you need on it. Disk space is only close to max because I haven't deleted old backups in awhile.


In need be, I can put you on this server;

Oh wow! I will probably do that! Thanks so much!

[xxpoopyxx]

I went to that site, what were you hosting there?

[Justin]

I went to that site, what were you hosting there?

Nothing, I only have the domain so I can access it without needing to remember the IP address

[Benny]

Looks good. when i get the chance i'll chuck a couple words at it.

[xxpoopyxx]

how did you get that domain anyway did you buy it or just host it yourself?

[xxpoopyxx]

Looks good. when i get the chance i'll chuck a couple words at it.

Sweet thanks!!

[xxpoopyxx]

VERSION 2 HAS BEEN RELEASED!!! Putting up a tutorial on main post soon!

[xxpoopyxx]

Still working on some PHP for keys and answers! having some problems! i should be finished with it tonight or tomorrow. Keep those words coming!!!

[rj]

Just voted no by accident.. can someone change it lol?

[xxpoopyxx]

haha thats alright! can you actually change them? i will if i can!

[rj]

Mods+ can I believe

[xxpoopyxx]

hmmm wellllll then we need to find one i guess

[xxpoopyxx]

Just added the keys and answers input for the website!! go try it!! adding tutorial next!!

[xxpoopyxx]

Just added the tutorial!! Come check it out!!!

[xxpoopyxx]

Le jingle and i have figured out more procedures that will be added to the second part of the script! wish us luck and help us out with words, keys, and answers!!

[Le Jingle]

Le jingle and i have figured out more procedures that will be added to the second part of the script! wish us luck and help us out with words, keys, and answers!!

Just sharing what knowledge I have to help
Will try to look at your project a little more over the next week too,

Cheers
Lj

[xxpoopyxx]

Just sharing what knowledge I have to help
Will try to look at your project a little more over the next week too,

Cheers
Lj

Thanks a bunch Le Jingle

[Sawyer]

I own the server so I can put whatever you need on it. Disk space is only close to max because I haven't deleted old backups in awhile.


In need be, I can put you on this server;

I wish my home internet had a 2 TB bandwidth...

[xxpoopyxx]

I admit I've never seen so much before!

Remember this script relies on the community! with out it, it wont go anywhere!

[Itankbots]

Amazing project man! I love how your keeping it community involved. Ill defintaly help add some stuff when i get free time, Currently very busy irl due to health issues, but i will defintaly find some time!

[cause]

Very very cool project. So, if I understand how your script will run:

Random Person: "Hello" [Hello -> index # is 45 for example]
Script -> server: [45] (will the server or the script pick the return index?)
server -> [returns one of: "Hi", "Hello yourself", or "Sup"]
Script -> types out the returned response.

A couple of thoughts:
Who is going to go through and index proper responses? Seems like it would take a long time. (If opened to the community, I guess we could tackle it fairly quickly)

Also, the response procedure should be fairly generic for easy implementation into scripts (obvious but thought I'd mention :P ). Hmmm, the response procedure should also be able to return skill levels I would imagine.

Great project overall!
(Also, poops, please don't double post )

EDIT: Also, to reduce server stress, maybe when the script sends a request to the server, it should include some type of randomization (random number between 1 and 9999 lets say, then add it up: ex. 435 = 4+3+5 = 12 then add up the return indexes, say, "Hi" => 923 = 9+2+3 = 14, "Hello yourself" => 243(hello) + 8421(yourself) = 8664 = 8+6+6+4 = 24, "Sup" => 78 = 7+8 = 15 THEN pick the number which is closest so, "Hi" when added up returns 14, so pick 14. This is just a suggestion to reduce load on the server, this would also mean that every word in your index should have its index number, and its index number added up. Hmmm well this is just a thought)

EDIT2: No, better idea, just send a random number between 0 and however long the index file is (145 000 words right?) and then just return the indexed result that is closest...wait no...this wouldn't work if words like "Hi" and "Hello" are closer together (since you probably added them close together). For this to work, you need to randomize the order of the words in the list...actually you could just setup a script to download the latest version to your computer, randomize it, then upload it again. On the other hand, I don't know how much stress picking a random number is on the server in which case it renders my whole train of thought null LOL.

EDIT3: Expanding on previous idea, you would not need to randomize the order of the list IF you assigned array numbers with a=1, b=2, c=3 etc... so "Hi" = 89 (h = 8, i = 9).

EDIT4: Also, potentially when you are indexing responses, assign different weights to the responses (a scale between 1-10?) So, give "Hi" a weight of 6, and "Hello yourself" a weight of 3, and "Sup" a weight of 4, then, add up the weights: 13, so "Hi" has a 6/13 chance of being said, where as "Sup" only a 4/13 chance. Since it seems that more often people say "Hi" instead of "Sup".

I was also thinking, we should try and make the word list a "Younger players" word list, it seems easier to pretend to be a younger player than an older player (potentially misspell harder words?)(hmmm maybe a separate procedure for determining if a word will be typed correctly or not)(this could also contribute to anti-ban, where you could specify at the beginning of the script (maybe a weighting process out of 10 again [maybe a 100]) that would determine how likely your player is to misspell words. (one person specifies a weight of 70, so their player is quite likely to misspell words vs a specification of say, 12, fairly unlikely to misspell a word.)(lol and to add to this, you could even have the script fix it's spelling mistakes:
Person: Hi
You: ho
You: *hi

that would look very human!!)

EDIT#watever: Does jagex still auto-capitalize and disallow all caps? I feel like there are numerous randomization options there

[xxpoopyxx]

Very very cool project. So, if I understand how your script will run:

Random Person: "Hello" [Hello -> index # is 45 for example]
Script -> server: [45] (will the server or the script pick the return index?)
server -> [returns one of: "Hi", "Hello yourself", or "Sup"]
Script -> types out the returned response.

A couple of thoughts:
Who is going to go through and index proper responses? Seems like it would take a long time. (If opened to the community, I guess we could tackle it fairly quickly)

Also, the response procedure should be fairly generic for easy implementation into scripts (obvious but thought I'd mention :P ). Hmmm, the response procedure should also be able to return skill levels I would imagine.

Great project overall!
(Also, poops, please don't double post )

EDIT: Also, to reduce server stress, maybe when the script sends a request to the server, it should include some type of randomization (random number between 1 and 9999 lets say, then add it up: ex. 435 = 4+3+5 = 12 then add up the return indexes, say, "Hi" => 923 = 9+2+3 = 14, "Hello yourself" => 243(hello) + 8421(yourself) = 8664 = 8+6+6+4 = 24, "Sup" => 78 = 7+8 = 15 THEN pick the number which is closest so, "Hi" when added up returns 14, so pick 14. This is just a suggestion to reduce load on the server, this would also mean that every word in your index should have its index number, and its index number added up. Hmmm well this is just a thought)

EDIT2: No, better idea, just send a random number between 0 and however long the index file is (145 000 words right?) and then just return the indexed result that is closest...wait no...this wouldn't work if words like "Hi" and "Hello" are closer together (since you probably added them close together). For this to work, you need to randomize the order of the words in the list...actually you could just setup a script to download the latest version to your computer, randomize it, then upload it again. On the other hand, I don't know how much stress picking a random number is on the server in which case it renders my whole train of thought null LOL.

EDIT3: Expanding on previous idea, you would not need to randomize the order of the list IF you assigned array numbers with a=1, b=2, c=3 etc... so "Hi" = 89 (h = 8, i = 9).

EDIT4: Also, potentially when you are indexing responses, assign different weights to the responses (a scale between 1-10?) So, give "Hi" a weight of 6, and "Hello yourself" a weight of 3, and "Sup" a weight of 4, then, add up the weights: 13, so "Hi" has a 6/13 chance of being said, where as "Sup" only a 4/13 chance. Since it seems that more often people say "Hi" instead of "Sup".

I was also thinking, we should try and make the word list a "Younger players" word list, it seems easier to pretend to be a younger player than an older player (potentially misspell harder words?)(hmmm maybe a separate procedure for determining if a word will be typed correctly or not)(this could also contribute to anti-ban, where you could specify at the beginning of the script (maybe a weighting process out of 10 again [maybe a 100]) that would determine how likely your player is to misspell words. (one person specifies a weight of 70, so their player is quite likely to misspell words vs a specification of say, 12, fairly unlikely to misspell a word.)(lol and to add to this, you could even have the script fix it's spelling mistakes:
Person: Hi
You: ho
You: *hi

that would look very human!!)

EDIT#watever: Does jagex still auto-capitalize and disallow all caps? I feel like there are numerous randomization options there

i recently talked with some people that gave me ideas on some better ways to process info so i will keep this way and add the other after i figoure out how it will work!

the way this works is basically they way you said except its reading from a text file to speed up the process not the server

it will take a sentence some one says and break that sentence into an array then go through each word in that array and compare it to a text file, after it compares each word it will fetch the index of each word and generate a key, after that the key is compared with others that have responses like hi, or hello.

so if the sentence is 'hi how are you'
then it will make an array ['hi', 'how', are', 'you']
then it will pick the first index of that array which is 'hi'
then it goes through a text file to find the word 'hi' then returns the index in the text file

so for example hi could equal the index of 2345

so it does that for each word and generates a key based on the indexes
Example:
2345-6789-12345-678

then that key is compared to another text file with a list of keys and if it finds that key in there then it will take a responses form that key

the keys look like this
2345-6789-12345-678 := ['im good', 'great', 'fine']

then it will pick a random response and say it

the new way that it should be done is with with relations built upon the sentences said, and having key words and all that other good stuff

the user 'core' has suggested that to me and agreed to help out when he can! so also thank core for helping, and the rest of the community for contributing

[cause]

i recently talked with some people that gave me ideas on some better ways to process info so i will keep this way and add the other after i figoure out how it will work!

the way this works is basically they way you said except its reading from a text file to speed up the process not the server

it will take a sentence some one says and break that sentence into an array then go through each word in that array and compare it to a text file, after it compares each word it will fetch the index of each word and generate a key, after that the key is compared with others that have responses like hi, or hello.

so if the sentence is 'hi how are you'
then it will make an array ['hi', 'how', are', 'you']
then it will pick the first index of that array which is 'hi'
then it goes through a text file to find the word 'hi' then returns the index in the text file

so for example hi could equal the index of 2345

so it does that for each word and generates a key based on the indexes
Example:
2345-6789-12345-678

then that key is compared to another text file with a list of keys and if it finds that key in there then it will take a responses form that key

the keys look like this
2345-6789-12345-678 := ['im good', 'great', 'fine']

then it will pick a random response and say it

the new way that it should be done is with with relations built upon the sentences said, and having key words and all that other good stuff

the user 'core' has suggested that to me and agreed to help out when he can! so also thank core for helping, and the rest of the community for contributing

Ok so are you building some type of sentence breakdown engine, or are you looking for predefined questions.
Like: "hi how are you", is both a greeting and a question...how would you decipher this?
What if I said: "whatsup, also, hi" ? That is definitely not going to be a pre-defined question.

[xxpoopyxx]

Ok so are you building some type of sentence breakdown engine, or are you looking for predefined questions.
Like: "hi how are you", is both a greeting and a question...how would you decipher this?
What if I said: "whatsup, also, hi" ? That is definitely not going to be a pre-defined question.

the beggining of this script is pre defined questions that are compared with an array of answer keys the better way to do this which i will is to use keywords and relations with a self generating database basically

[cause]

the beggining of this script is pre defined questions that are compared with an array of answer keys the better way to do this which i will is to use keywords and relations with a self generating database basically

Ok I see. Once you have a delivery system made I can help with randomization.