[Feature] Default max_prompt_tokens should be LLM dependent

[igiloh-pinecone]

Is this your first time submitting a feature request?

• I have searched the existing issues, and I could not find an existing issue for this feature
• I am requesting a straightforward extension of existing functionality

Describe the feature

The ChatEngine takes a max_prompt_tokens argument, which allows limiting the prompt size (message history + system prompt + retrieved context).
Currently, this argument has an arbitrary hard-coded default value of 4096. However - the real default value should be LLM dependent, as different LLMs have different context windows.

Describe alternatives you've considered

Each LLM class should have a context_window cached property (the value may change, according to model_name).
If the user hasn't specified max_prompt_tokens (the default should be None) - the ChatEngine will use llm.context_window.

Who will this benefit?

No response

Are you interested in contributing this feature?

No response

Anything else?

No response