[chore] support tool calling (#59)

* allow --jinja flag

* docs

* correct handling of is_function_calling

llama-index embedding (#58)

* llama-index embedding

* mypy overrides

ALI-1427 handle windows ai nav config path (#54)

* handle windows ai nav config path

* separate windows logic

* revert separation of platform logic

* pin click to <8.2 to fix python 3.10+ errors

Vectordb methods (#51)

* list-tables, create-table, drop-table

* raise_for_status clean up

[chore] config, cli, params refimenent (#31)

* refine cli, config, and param support

* switch integrations to modelFileName

* refresh notebooks

* download success

* timeout the download start delay

parse tensor-split as list of numbers (#27)

[feat] improve parameter support in integrations (#26)

* langchain: parameter support

* server params with litellm

* llama-index

* drop params from notebook

* panel kwargs

* fix litellm + dspy

* server options for llm

* mypy ignores

* two more

[chore] fix stop on exit (#25)

* internalize config support

* remove atexit from integrations

* update examples

* config in readme

* ignore **kwargs usage

[chore] docs (#24)

* docs

* action

[chore] fix llm and langchain (#14)

* almost all variants for infer and embed

* add pooling for embedding for safety

* incorrect value for langchain llama-cpp options