vLLM Provider lacks v1 url

[hbai98]

When setting
export VLLM_API_BASE="http://0.0.0.0:8888" # Run the Letta server (usually on a different port) poetry run letta server --host 0.0.0.0 --port 80 # Add --ade if needed for development

or export VLLM_API_BASE="http://localhost:8283, the vLLM model can't be found.

To resolve it, adding a suffix like 'export VLLM_API_BASE="http://0.0.0.0:8888/v1', benefits finding the models available and latter functions' calls.

It may be helpful to set up like:

letta/letta/server/server.py

Lines 313 to 319 in 79f4a7e

	# Attempt to enable LM Studio by default
	if model_settings.lmstudio_base_url:
	# Auto-append v1 to the base URL
	lmstudio_url = (
	model_settings.lmstudio_base_url
	if model_settings.lmstudio_base_url.endswith("/v1")
	else model_settings.lmstudio_base_url + "/v1"

For adding this suffix automatically.