feat: added with statement support to release memory and exposed help function for tokenizer

[guocuimi]

• added context management for LLM and Engine

  with AsyncLLMEngine(...) as engine: 
  engine.schedule(...)

• added help function for tokenizer

  def apply_chat_template(self, messages: List[Message]) -> Optional[str]: ...
  def encode(self, text: str) -> List[int]: ...
  def decode(self, tokens: List[int], skip_special_tokens: bool) -> str: ...