extract OpenAI client as a separate module & add `max_tokens` setting

fix lint issues

add supports for conversation mode

refactor: unified command handling & dispatching chain

remove latest branch from docker files

fix wording

add instructions for Docker

update readme

add "Show Raw Contents" button for Markdown message

fallback to raw contents when failed to send rendered Markdown

fix wrong Markdown rendering result when input is only one code block

isolate parser errors to avoid panics

finish the Markdown feature & add config

optimize code block rendering

fix linebreak issues in Markdown text