Enable MiniMax chat completions configuration

[Octopus (octo-patch)]

Reason: Enable MiniMax through configurable chat completions endpoints and current model IDs.

• Read base_url and model from openaiapirc, with the legacy engine key retained as a compatibility alias.
• Use the chat completions client and remove the obsolete secondary completion request.
• Update Bash, Zsh, and PowerShell setup flows plus documentation for the global and China endpoints.
• Add focused tests for both regional endpoints, current models, chat requests, and saved configuration.

Checks:

• python -m unittest discover -s tests -v
• python -m compileall -q src tests
• bash -n scripts/bash_setup.sh
• zsh -n scripts/zsh_setup.sh
• git diff --check