Enable Langserve chat server and playground

[virajmalia]

The new architecture looks something like this:

(1) Ollama instance hosted locally (localhost:8080) --> (2) add tools and agents with langchain \
--> (3) host LLM chain locally using langserve(localhost:8000)

This architecture allows development and testing of (2) (3), while making sure the output is consistent or better than the native output from the LLM chat model from Ollama.
(1) also hides a lot of configuration required to host a chat model. This may change in the future.