artero / jambots Goto Github PK

Now this value is hardcoded in the code:

# lib/jambots/bot.rb
@client = OpenAI::Client.new(access_token: openai_api_key, request_timeout: 240)

The openai_api_key is utilized in the Bot#initialize method, but it is not used in the controllers. Allowing this to be added as a bot option could prove useful.

Introducing a minimalist renderer without extra information could be beneficial for using the jambots chat command within a pipe. This would streamline the output and make it more suitable for further processing in such scenarios.

exe/jambots chat # Uses CLI
exe/jambots chat --format plain # no pretty
exe/jambots chat --format json # Some kind of json thing

And --no-format, -n as shortcut of --format plain

from this conversation #9 (comment)

Add an examples_and_experiments directory and include an example demonstrating how I'm using Jambots for development in RoR.

Allow the possibility to use llama.cpp models.

Planned changes:

• We need to add new options in the bot.yml file to specify which client the bot should run with and any extra client configurations required by the clients.
• We should create a factory that enables the use of different clients.
• Develop OpenAI and LLamaCpp client implementations.

Implement a custom Error for cases when the OpenAI API returns an error response. In the "ChatController," we can catch and display a more descriptive error message to the user via stderr.