feat: add FunASR speech-to-text provider

[LauraGPT]

Summary

• add a dedicated FunASR speech-to-text provider to the existing chatflow STT configuration
• default to the local OpenAI-compatible endpoint http://127.0.0.1:8000/v1 with model sensevoice
• support optional language, base URL, model, and gateway API-key configuration
• keep existing OpenAI and LocalAI saved configurations unchanged and add no new runtime dependency

Closes #6454.

Validation

• TDD red/green coverage for no-credential local defaults and optional gateway settings
• pnpm --filter flowise-components test --runInBand: 21 suites, 876 tests passed
• focused FunASR tests: 2 passed
• changed-file ESLint: passed with zero warnings
• pnpm --filter flowise-components build: passed
• pnpm --filter flowise-ui build: passed
• real CPU smoke test against FunASR 1.3.26 using the same @langchain/openai client path: health 200, transcription 200, and a valid transcript returned

[gemini-code-assist]

Caution

The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased.

[LauraGPT]

Refreshed this PR onto current FlowiseAI/main@ababf4c8; the exact signed head is now 1ab8f4d3668c64e75df635e8a36fb25972c34d68.

Fresh exact-head verification:

• all 21 component test suites pass, 892/892 tests total, including both FunASR paths
• changed-file ESLint passes with zero warnings
• flowise-components TypeScript/Gulp production build passes
• flowise-ui Vite production build passes
• ED25519 commit verification and git diff --check pass
• GitHub reports the PR mergeable with no unresolved review threads

The provider remains dependency-free and preserves existing OpenAI/LocalAI settings while adding local FunASR defaults plus optional gateway configuration. @HenryHengZJ, could you review this speech-to-text integration when convenient?