Add Baidu transcription provider

[mitre88]

Summary

• add a Baidu Speech Recognition provider behind --api baidu
• use environment-based BAIDU_API_KEY / BAIDU_SECRET_KEY configuration with no committed secrets
• convert Baidu audio input to mono 16 kHz MP3 and send the short speech recognition request through Baidu's token + server_api flow
• add focused unit coverage for token usage, request payload construction, language model routing, missing credentials, and API errors
• document the new environment variables and CLI option

Validation

• .venv/bin/python -m unittest discover -s tests -v
• .venv/bin/python -m compileall src tests
• .venv/bin/sapat --help
• git diff --check

No secrets, payout details, or private audio are included.