Add Markdown Support & Translator Model

[fuma-nama]

Goal

Support Markdown, this includes support for handling diffs.

Changes

• Redesign the current translate command
  • Add a layer of Translator
  • Two types of operation: update and new. I found the differences between the two operations could be large and no longer can be grouped by a single force: true.
  • Re-implemented the current translators (e.g. json, js/ts)
• Added unit tests with Vitest
  • ensure the previous translators still work
  • do not require a token, contributor-friendly

Context

Usually we want to maintain the context of docs, it would be better to provide the full content.

This PR will provide the full-context, which needs more input tokens. Maybe we think of a better algorithm(?)

(I ran bun format, some previous files are somehow updated too.)

Diff

the current diff command cannot detect the real updates, translator handled the diff step. We can probably make it work with the new translator model by another PR

[vercel]

@fuma-nama is attempting to deploy a commit to the Pontus Abrahamsson's projects Team on Vercel.

A member of the Team first needs to authorize it.

[fuma-nama]

I changed the current implementation:

• JSON translator will compare the two objects instead of using git diff, it should be more accurate.
• JS/TS translator directly replace strings, like:

const v = "this will be translated"

We assume the order of strings will not be changed during translation, so it can be cached by reusing previous translations

I have added some details via code comments to the new algorithm.

[fuma-nama]

Tested only with o4 on unit tests, might need your help to test it with other models and on CLI