This PR makes several large internal changes to support the ability to use templating engines other than OpenAI models. Notably, we now support running with local LLM models (llama3 and llama3-70b) thru Ollama. A new common abstraction of "Renderer" has been introduced to support different rendering pathways.
Breaking changes
--model/-m flag is now deprecated in favor of the new --renderer/-r flag.
A list of supported renderers can be found using the new command mailmerge renderers list
Description
This PR makes several large internal changes to support the ability to use templating engines other than OpenAI models. Notably, we now support running with local LLM models (llama3 and llama3-70b) thru Ollama. A new common abstraction of "Renderer" has been introduced to support different rendering pathways.
Breaking changes
--model/-m
flag is now deprecated in favor of the new--renderer/-r
flag.A list of supported renderers can be found using the new command
mailmerge renderers list