AI Bots 3.0.2

[021]

021 updated [021] AI Bots with a new update entry:

1.1.0

Updated minimum requirements
ChatGPT Framework 1.4.0+

GPT-4 models support for ChatGPT and GPT Dall-E both
You will be able to use this model only if you have been granted access to it.

For the ChatGPT handler in the Tuning tab, settings for different contexts are now in collapsible blocks
This will improve convenience.

For developers
Added service for streaming bot responses from ChatGPT.

Read the rest of this update entry...

 

[021]

Does this contribute to each request's token amount? For example, if the last 10 messages have 8k tokens, will some of that be truncated, and will that impact or limit the reply / output from the bot?

Yes, it does. No, messages will not be truncated.

If I were to put "100" in that field, and the text in a thread's replies total 50k tokens, does the bot only use the most recent posts, does it reject the request, or something else?

You won't be able to set this value as the maximum number for this field is limited so that the bot doesn't run into a token limit issue. This value increases depending on the chat model.

Similarly, does the bot "prompt in threads" text also contribute/count towards the tokens input and limit?

Yes

 

[ProWeb]

Thanks for the replies!

No, messages will not be truncated.

So, what happens in this scenario:

The "Context limit in threads" is set to 5.
The quoted message is 5k tokens and each of the previous 5 messages are also 5k tokens each.

What does the bot do in that situation where every input is over the token limit?

 

[ProWeb]

Thanks for the replies!

So, what happens in this scenario:

The "Context limit in threads" is set to 5.
The quoted message is 5k tokens and each of the previous 5 messages are also 5k tokens each.

What does the bot do in that situation where every input is over the token limit?

I think I just found the (frustrating) answer.

We're not getting replies to the chatbot... just like it's dead.

I checked error-log and found this:


"This model's maximum context length is 4097 tokens. However, your messages resulted in 4213 tokens. Please reduce the length of the messages.",
"type": "invalid_request_error",

Bummer there isn't a way around this so that either the data sent isn't "too long" or the bot returns some friendly error in the reply like "That was too long, can you try again or contact the admin"

 

[Lee]

I am struggling to set up multiple bots.

I have one set up like this:





This bot is supposed to scan all messages for suicidal content and respond with help lines. It does this successfully.

But then I have a bot configured like this:



So why did the first bot post the reply to a technical question, not the second bot? How do I change my avatar? | The Life Spot



This resulted in a loop of the bots replying to themselves.

 

[021]

So why did the first bot post the reply to a technical question, not the second bot? How do I change my avatar? | The Life Spot

Because both bots are set to respond to all messages in the selected section, and TLS Bot is the first one to do so.
You can enable the Smart ignore option and add something like this to the prompt:
(not tested to work)

Parse user message for suicidal content.
Ignore user message if there is no suicidal content.

Thus, the bot should ignore messages that do not contain suicidal content.

This resulted in a loop of the bots replying to themselves.

It's funny) It turns out that bots can communicate with each other.
I will add an option so that bots can ignore other bots.

 

[kprojects]

I'd like the bot to only reply in threads if he's mentioned.. tried different settings but can't figure it out?

 

[021]

I'd like the bot to only reply in threads if he's mentioned.. tried different settings but can't figure it out?

Hello. Follow the instructions from the guide, at the stage of setting up triggers, disable the post quote trigger.

 

[kprojects]

Hello. Follow the instructions from the guide, at the stage of setting up triggers, disable the post quote trigger.

That's how I currently have it set up, however, he still replies in the thread w/o being mentioned.

 

[021]

That's how I currently have it set up, however, he still replies in the thread w/o being mentioned.

You probably set up a Posted in trigger - you don't need it.

 

[kprojects]

You probably set up a Posted in trigger - you don't need it.

Yes .. thanks!

 

[021]

021 updated [021] AI Bots with a new update entry:

1.2.0

Important ℹ️​

This version requires the installed add-on [021] ChatGPT Framework 1.5.0+

Bot profile handling 🙆‍♂️​

Bots can now reply to posts (and comments) on their profile – the perfect place to communicate with AI.
New permissions have also been added:
In connection with this innovation, new permissions have been added for both handlers:
Use bots in their profiles – Allow user to receive replies from bots in their profiles.
Max replies in bot profile per day – Limits the maximum number of replies for a user in a bot profile.

[ATTACH...

Read the rest of this update entry...

 

[snerd]

In my ACP, under Forums, there is a listing for ChatGPT node prompts. There is nothing listed in that area. What's it for?

 

[kprojects]

Noticed today after installing v1.10 that when mentioning someone, it's always @user now instead of their actual username.

 

[rdn]

Now when the bot responds, its latest activity is updated and the online status is also displayed.

Please also add this on autoresponder addon :)

 

[021]

In my ACP, under Forums, there is a listing for ChatGPT node prompts. There is nothing listed in that area. What's it for?

In this section, you can set up an individual prompt for each forum section for bots built on the ChatGPT handler.
For example, you have a section where users ask questions about a specific software that you release - you can provide the necessary information about this software and configure the bot to answer questions in this section in a specific way. While other sections will have a global prompt for threads, which will save the use of tokens.

Noticed today after installing v1.10 that when mentioning someone, it's always @user now instead of their actual username.

This update did not make any changes that could affect this behavior.
By default, ChatGPT does not know users' nicknames, but it can be taught this by adding to prompt something like:
You can mention the member with @{author}
The add-on practically does not regulate the behavior of bots.

 

[WesSDG]

Hello there.
I've installed the framework and provided the API key. However when I install the AI Bots addon I get this error below. Anyone have any suggestions?

ParseError: syntax error, unexpected 'Client' (T_STRING), expecting function (T_FUNCTION) or const (T_CONST) in src/addons/BS/AIBots/Service/LicenseChecker.php at line 13

1. Composer\Autoload\includeFile() in src/vendor/composer/ClassLoader.php at line 428
2. Composer\Autoload\ClassLoader->loadClass()
3. spl_autoload_call()
4. class_exists() in src/XF/Container.php at line 267
5. XF\Container->createObject() in src/XF/App.php at line 1725
6. XF\App->XF\{closure}() in src/XF/Container.php at line 234
7. XF\Container->create() in src/XF/App.php at line 3158
8. XF\App->service() in src/XF.php at line 901
9. XF::service() in src/addons/BS/AIBots/Setup.php at line 37
10. BS\AIBots\Setup->checkRequirements() in src/XF/AddOn/AddOn.php at line 457
11. XF\AddOn\AddOn->checkRequirements() in src/XF/Job/AddOnInstallBatch.php at line 236
12. XF\Job\AddOnInstallBatch->stepPreAction() in src/XF/Job/AddOnInstallBatch.php at line 87
13. XF\Job\AddOnInstallBatch->run() in src/XF/Job/Manager.php at line 260
14. XF\Job\Manager->runJobInternal() in src/XF/Job/Manager.php at line 202
15. XF\Job\Manager->runJobEntry() in src/XF/Job/Manager.php at line 118
16. XF\Job\Manager->runByIds() in src/XF/Admin/Controller/Tools.php at line 122
17. XF\Admin\Controller\Tools->actionRunJob() in src/XF/Mvc/Dispatcher.php at line 352
18. XF\Mvc\Dispatcher->dispatchClass() in src/XF/Mvc/Dispatcher.php at line 263
19. XF\Mvc\Dispatcher->dispatchFromMatch() in src/XF/Mvc/Dispatcher.php at line 115
20. XF\Mvc\Dispatcher->dispatchLoop() in src/XF/Mvc/Dispatcher.php at line 57
21. XF\Mvc\Dispatcher->run() in src/XF/App.php at line 2483
22. XF\App->run() in src/XF.php at line 524
23. XF::runApp() in admin.php at line 13

 

[021]

Hello there.
I've installed the framework and provided the API key. However when I install the AI Bots addon I get this error below. Anyone have any suggestions?

ParseError: syntax error, unexpected 'Client' (T_STRING), expecting function (T_FUNCTION) or const (T_CONST) in src/addons/BS/AIBots/Service/LicenseChecker.php at line 13

1. Composer\Autoload\includeFile() in src/vendor/composer/ClassLoader.php at line 428
2. Composer\Autoload\ClassLoader->loadClass()
3. spl_autoload_call()
4. class_exists() in src/XF/Container.php at line 267
5. XF\Container->createObject() in src/XF/App.php at line 1725
6. XF\App->XF\{closure}() in src/XF/Container.php at line 234
7. XF\Container->create() in src/XF/App.php at line 3158
8. XF\App->service() in src/XF.php at line 901
9. XF::service() in src/addons/BS/AIBots/Setup.php at line 37
10. BS\AIBots\Setup->checkRequirements() in src/XF/AddOn/AddOn.php at line 457
11. XF\AddOn\AddOn->checkRequirements() in src/XF/Job/AddOnInstallBatch.php at line 236
12. XF\Job\AddOnInstallBatch->stepPreAction() in src/XF/Job/AddOnInstallBatch.php at line 87
13. XF\Job\AddOnInstallBatch->run() in src/XF/Job/Manager.php at line 260
14. XF\Job\Manager->runJobInternal() in src/XF/Job/Manager.php at line 202
15. XF\Job\Manager->runJobEntry() in src/XF/Job/Manager.php at line 118
16. XF\Job\Manager->runByIds() in src/XF/Admin/Controller/Tools.php at line 122
17. XF\Admin\Controller\Tools->actionRunJob() in src/XF/Mvc/Dispatcher.php at line 352
18. XF\Mvc\Dispatcher->dispatchClass() in src/XF/Mvc/Dispatcher.php at line 263
19. XF\Mvc\Dispatcher->dispatchFromMatch() in src/XF/Mvc/Dispatcher.php at line 115
20. XF\Mvc\Dispatcher->dispatchLoop() in src/XF/Mvc/Dispatcher.php at line 57
21. XF\Mvc\Dispatcher->run() in src/XF/App.php at line 2483
22. XF\App->run() in src/XF.php at line 524
23. XF::runApp() in admin.php at line 13

Hello. Minimum PHP version for this add-on to work is 7.4

 

[WesSDG]

Hello. Minimum PHP version for this add-on to work is PHP 7.4

We just figured that out. Thanks for the quick note.

 

[akok]

Can an autoresponder be added in case users exceed the token limit and the bot stops responding? Is it possible to notify the user that the token limit has been exceeded? Also, as an idea, could a button be added to prevent sending context older than a specific post?