tgpt -f does not work, history does not work on non interactive modes either.
xplshn opened this issue · comments
[anto@beef ~]$ tgpt -f
Usage: tgpt [Flag] [Prompt]
Flags:
-s, --shell Generate and Execute shell commands. (Experimental)
-c, --code Generate Code. (Experimental)
-q, --quiet Gives response back without loading animation
-w, --whole Gives response back as a whole text
-img, --image Generate images from text
Options:
-v, --version Print version
-h, --help Print help message
-i, --interactive Start normal interactive mode
-m, --multiline Start multi-line interactive mode
-cl, --changelog See changelog of versions
-u, --update Update program
Examples:
tgpt "What is internet?"
tgpt -f
tgpt -m
tgpt -s "How to update my system?"
[anto@beef ~]$
There is no tgpt -f and history is only in interactive modes by choice
see tgpt --help. tgpt used to save history on non interactive modes, and -f used to change the chat id... Why was that removed? It was pretty cool... It allowed me to do scripts using tgpt.
t[anto@beef ~]$ tgpt --help
Usage: tgpt [Flag] [Prompt]
Flags:
-s, --shell Generate and Execute shell commands. (Experimental)
-c, --code Generate Code. (Experimental)
-q, --quiet Gives response back without loading animation
-w, --whole Gives response back as a whole text
-img, --image Generate images from text
Options:
-v, --version Print version
-h, --help Print help message
-i, --interactive Start normal interactive mode
-m, --multiline Start multi-line interactive mode
-cl, --changelog See changelog of versions
-u, --update Update program
Examples:
tgpt "What is internet?"
tgpt -f # I mean, it is still listed as an example. I thought it was still a feature. It was comfortable to use that way.
tgpt -m
tgpt -s "How to update my system?"
[anto@beef ~]$
Yeah I have to remove it from examples.
Because the mechanism has changed. Before, it just stored a chatid, now it stores messages
How to persist the chat history between subsequent invocations of tgpt? Or how to append prompts to the interactive mode?
Something like tgpt -i < prompts.txt should be possible, with subsequent prompts on separate lines in prompts.txt.
I really liked that feature. Maybe could come back as a flag? Like --remember/-r ?
With the introduction of multiple API's it will be tougher to implement
Will have to use a cache file that holds the conversations the user's had woth every model, and tgpt should use the last used model