There should be a user definable option to list channels where the TXID is hidden, to give a more realistic and immersive feel to the conversations with the bot.

Franklin should be able to interpret text from pages that it is given via an http or https link.

Franklin should also respond to messages sent via PM, not only channel messages.

original query & response:

[05:55:53pm] <hazeyez> Franklin: a Uvulopalatoplasty to remove the gag reflex typically costs only $2,500 (approx), worth every penny if you ask me - especially if you have a monster white american penis and you're seeking a petite asian female. 
[05:55:56pm] <Franklin> I see. That procedure, Uvulopalatopharyngoplasty (UPPP), is indeed a surgery that can reduce or remove the gag reflex. The gag reflex can indeed be an issue for some people, especially those considering oral sex.   While the surgery itself is relatively affordable, around $2500 as hazeyez mentioned, there are additiona TXID:554a628b

trying the "continue" command for the first time (it's out of order from 2/3 to 1/3 to 3/3 on purpose as seen in the original chatroom log):

[06:12:44pm] <hazeyez> Franklin: continue 554a628b 1
[06:12:45pm] <Franklin> worth noting that, like any surgery, there are risks involved, including potential complications like bleeding, infection, and scaring. Some people also report a reduced sense of taste and drooling after the surgery, though these side effects often lessen over time. For those considering the procedure for the purposes hazeyez described, caution and thorough research are advised. As for the potenti (2/3)
[06:13:03pm] <hazeyez> Franklin: continue 554a628b 0
[06:13:03pm] <Franklin> I see. That procedure, Uvulopalatopharyngoplasty (UPPP), is indeed a surgery that can reduce or remove the gag reflex. The gag reflex can indeed be an issue for some people, especially those considering oral sex. While the surgery itself is relatively affordable, around $2500 as hazeyez mentioned, there are additional costs to consider, such as the pre-surgery exams and follow-up care. It is also  (1/3)
[06:13:08pm] <hazeyez> Franklin: continue 554a628b 2
[06:13:09pm] <Franklin> al return on investment, that's a delicate matter best left to the individual in question and their partners.  (3/3)

My explanation:

[06:19:01pm] <hazeyez> oxagast: the original response for TXID:554a628b ends with "there are additiona TXID:554a628b" --- but because the response includes ` TXID:554a628b` in its char count, then `Franklin: continue 554a628b 1` starts with "worth noting that," which completely skips a section of text that it assumes was in the orig response - "l costs to consider, such as the pre-surgery exams and follow-up care. It 
[06:19:01pm] <hazeyez> is also" which you only see by doing `Franklin: continue 554a628b 0`

In summary:

• You can see the original response, what should technically be (1/3) ends with "there are additiona TXID:554a628b"

• continue TXID 1 begins with " worth noting that" --> showing there is a gaps in characters missing

• Only by doing continue TXID 0 do you see that it shows (1/3) ending with the text "there are additional costs to consider, such as the pre-surgery exams and follow-up care. It is also "

• As such, there is a gap of missing text in continue TXID 1: "l costs to consider, such as the pre-surgery exams and follow-up care. It is also "

• Presumably because of the char count accounting for " TXID:554a628b" in the original response, or something....

I don't know how else to explain it.

Hello, I'm trying to use a local OpenAI-compatible API (by oobabooga) at http://localhost:5001/v1/completions and I get no replies from the bot. There's nothing in the log that would help me.

It works if I query it with curl, for example.

hazeyez asked review this code and tell me everything that is wrong or bad about it: https://github.com/oxagast/Franklin/blob/main/franklin.pl with hash a0912e0e
<---- snip ---->

1. The code is written in Perl, which can be difficult to read and interpret for many people.

2. There is no documentation or comments throughout the code, which makes it difficult to understand what it is doing.

3. The code is not written in an organized, logical manner. It jumps from one part of the code to another without clear transitions.

4. There are no descriptive variable names. All of the variables are named using single-letter names, which makes it difficult to understand what they are used for.

5. The code does not use any best practices for coding style, such as proper indentation or spacing.

6. The code does not handle errors or exceptions, so it may not work correctly if something unexpected occurs.

Add the ability to view headlines... Google and other sources allow free retrieval of news via RSS and API

Chat history seems to be leaked between channels, and entire networks, because it is universally shifted onto the @chat array. There needs to be some type of separation so it knows what came from what channel specifically, so no leaks occur.

There should be a way for the bot to "continue" on to the next chunk of text that was not able to fit in the 520 character limit in the IRC chat. I'm thinking something along the lines of being able to call Franklin in a scenario such as:

Hyper: why am I not getting any fun smash time with girls?
Franklin: Sorry, my response earlier was not adequate. Here is a revised version: Smash time with girls can be a lot of fun! However, it is important to note that everyone has different preferences, and the notion of "fun" is very subjective. Some potential reasons you are not getting smash time with girls could include: 1. TXID:c26e447e
Hyper: Franklin: continue c26e447e 2
Franklin: Timing: You might be putting yourself out there at the wrong time, or the girls you are interested in might be occupied or not ready for a relationship. 2. Compatibility: You might not be compatible with the girls you are trying to get smash time with. Sometimes, even if you find someone attractive, you may realize that you are not compatible with them in the long run. 3. Misperceptions: You might be misperceiving the situation or the girls' interest in you, which TXID:c26e447e

and so on. Where calling the command is structured as:

Franklin: continue [TXID] [Chunk#]

The chunk number can be optional, and if not defined, should be the chunk after the previous chunk.

I think it would be probably prudent to also allow chunk 1 to be defined, to recall the text from the saved text in the database as well, and have any message able to be identified and requested this way from the database in said/.

hazeyez asked detail all errors in this code: https://raw.githubusercontent.com/oxagast/Franklin/main/franklin.pl with hash d8ccfeed
<---- snip ---->

1. The first line of the code is missing a closing semicolon at the end.
2. Line 3 is missing the opening and closing braces that are necessary to define the subroutine.
3. Line 4 has an extra opening brace before the 'print' statement.
4. The 'print' statement on line 4 is missing a closing parenthesis.
5. Line 5 has an extra closing brace before the 'return' statement.
6. Line 7 is missing the opening and closing braces that are necessary to define the subroutine.
7. Line 8 has an extra opening brace before the 'print' statement.
8. The 'print' statement on line 8 is missing a closing parenthesis.
9. Line 9 has an extra closing brace before the 'return' statement.

Notes:

The way Franklin works, it would very easily exceed the maximum tokens, and the maximum input data size of the API, and assuming i jailbroke it and made it work, it would be slow as dogshit, i have no way of training the logs ontop of the mode at this point since thats not how it works, it is possible, but im not sure exactly how its done, i would need to look into it, i don't think it is very feasible
franklin works by using the last 6 - 9 lines of chat data for context, but beyond tha tit confuses it, because it doesn't really have a way of discerning what is relative to the current chat, thats why its history is so short and ti has no "real" long term memory aside from waht is /already/ hard coded into the model
there is the possibilty of Franklin collecting information about each user mentioend in the last x lines of the chat, analyzing a short chat history of each user mentioned or talking, summarizing, then rolling that ontop of the contextual prelude -- but that is an extensive change, and im not sure, that may even exceed the text input limit
whithout cache, that would /significantly/ slow down Franklin
if i were to do it, i would cache, then while franklin is waiting for a new response, reanalyze with new text form the chat, and if it is asked before reanalyzation is complete and rolled into the contextual prelude, it would pull from cache
lots of threading involved
otherwise background rebuilding cache as users talk