How to request translation of a very long document to LangChain?

[9mean2]

The documents may be very long, resulting in errors exceeding tokens per minute

If I ask to translate a long document at the LangChain prompt, how can I do so efficiently?

I even went so far as to separate it into chunks. The current logic is hard-coded to enter prompts one by one from doc[0] to doc[5].

Is there a way to do this efficiently?

I would appreciate it if you could also provide example code.