You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When closing tenere, the current chat will not be saved.
Now that history is persistent and ctrl + s is no longer used to manually save the chats, it would make sense to make sure the program saves the current chat before exiting the program, such as when pressing "q".
Otherwise, the user has to start a new chat before closing to ensure the previous chat is saved.
Or bring back the ctrl + s keybinding that will cause the current tenere.archive-% to be saved.
O perhaps do both; bring back the ctrl + s keybinding so user can have peace of mind their chat was saved... then call that function whenever program goes for shutdown.
BTW thanks for making the history persistent, works well (I did notice a history file being overwritten somehow though). I had a related idea in mind when suggesting persistent history that would support a kind of projects functionality. Basically would work how it does currently but instead of chats being saved to .local/share/tenere/tenere.archive-%, there would be a subdirectory for each project with the relative chats such as .local/share/tenere/my-project/tenere.archive-%.
One could either select the project in the terminal somehow or just specify it via command line, then whatever is in the history is related to that specific project. If no project is specified it could just use a default project directory where the tenere.archive-% files get dumped which could be .local/share/tenere/default/
tenere -p my-project would create a ./local/share/tenere/my-project directory if it doesn't exist,
persistent history functionality would then save tenere.archive files there
tenere -p my-project if ./local/share/tenere/my-project already exists would create the history based on files in ./local/share/tenere/my-project instead of default data directory
and if one simply runs tenere without specifying project if loads history from ./local/share/tenere/default/
Additionally, if one is specifying a project when loading tenere, loading the last saved chat for that project may make sense, but doesn't make sense when starting tenere without specifying a project. That being said I don't think this would be an option until an entirely different feature is built which would allow for the continuance of previous chats after they have been closed ( by opening history, pressing enter on an entry and that loads that chat into main view where the chat can be continued ).
I think having this kind of projects functionality would be a super useful feature for heavy LLM users.
Also, given the simplicity of tenere's UI it would be very down to earth for cmd line users.
The text was updated successfully, but these errors were encountered:
When closing tenere, the current chat will not be saved.
Now that history is persistent and ctrl + s is no longer used to manually save the chats, it would make sense to make sure the program saves the current chat before exiting the program, such as when pressing "q".
Otherwise, the user has to start a new chat before closing to ensure the previous chat is saved.
Or bring back the ctrl + s keybinding that will cause the current tenere.archive-% to be saved.
O perhaps do both; bring back the ctrl + s keybinding so user can have peace of mind their chat was saved... then call that function whenever program goes for shutdown.
BTW thanks for making the history persistent, works well (I did notice a history file being overwritten somehow though). I had a related idea in mind when suggesting persistent history that would support a kind of projects functionality. Basically would work how it does currently but instead of chats being saved to
.local/share/tenere/tenere.archive-%
, there would be a subdirectory for each project with the relative chats such as.local/share/tenere/my-project/tenere.archive-%
.One could either select the project in the terminal somehow or just specify it via command line, then whatever is in the history is related to that specific project. If no project is specified it could just use a default project directory where the
tenere.archive-%
files get dumped which could be.local/share/tenere/default/
tenere -p my-project
would create a./local/share/tenere/my-project
directory if it doesn't exist,persistent history functionality would then save
tenere.archive
files theretenere -p my-project
if ./local/share/tenere/my-project already exists would create the history based on files in./local/share/tenere/my-project
instead of default data directoryand if one simply runs
tenere
without specifying project if loads history from./local/share/tenere/default/
Additionally, if one is specifying a project when loading tenere, loading the last saved chat for that project may make sense, but doesn't make sense when starting tenere without specifying a project. That being said I don't think this would be an option until an entirely different feature is built which would allow for the continuance of previous chats after they have been closed ( by opening history, pressing enter on an entry and that loads that chat into main view where the chat can be continued ).
I think having this kind of projects functionality would be a super useful feature for heavy LLM users.
Also, given the simplicity of tenere's UI it would be very down to earth for cmd line users.
The text was updated successfully, but these errors were encountered: