Optimize for MTTRE: Mean Time To Running Examples

[codekiln]

It's awesome there are so many examples for langserve.

As someone trying to get up to speed quickly on the relationship between LCEL input / output types and it's relationship to the APIs and the playground, I want to be able to clone langserve and run any given example in less than 10 min.

Things that would help me get up and running faster

Make it possible to easily install all the dependencies of the examples

See also Feature Request: run poetry install including dependencies of the examples · Issue #678 · langchain-ai/langserve.

Bonus points: include a Dockerfile and associated makefile commands, so I don't even need to worry about python virtualenvs or python versions.

Include a built-in way to configure and supply environment variables for all the examples at once

A low-lift way would be to expect a developer to supply a .env file and use python-dotenv.

Include a built-in way to be able to change from one LLM to another

People have access to different LLMs. To decrease the needed churn, provide some way to flip out LLMs, so they can use the one they already have API keys for (and that's not always OpenAI!).