CLI Enhancement: Add Environment Variable Support for Model Name and API_KEY

[lemisky]

PR Message

This PR adds support for two environment variables to improve MinerU’s compatibility when deployed on model management platforms such as GPUStack.

Background

We are from the GPUStack team. While integrating MinerU into model management platforms, we encountered the following issues:

1. Unable to specify model_name when multiple models are present
   When MinerU calls /v1/models, it raises an error if the backend returns more than one model, as shown in the current implementation.
   The CLI does not provide a way to specify the model name.
   The new MINERU_VL_MODEL_NAME environment variable addresses this limitation.

2. Lack of API key support for authenticated platforms
   Many model management platforms use OpenAI-compatible API keys for authentication, but MinerU currently does not support providing an API key.
   The new MINERU_VL_API_KEY environment variable enables MinerU to work in authentication-required environments.

Changes

• Added support for MINERU_VL_MODEL_NAME to explicitly specify the model name.
• Added support for MINERU_VL_API_KEY to allow MinerU to pass an API key when authentication is needed.

These enhancements make MinerU more flexible and better suited for deployment behind AI gateways and model management platforms that handle multiple models or require authentication.

[lemisky]

@myhloli , could you please take a look at this PR when you have time? Thanks!