toor
1e9ea89d86
Import the client, server, skills, tests, and build scripts as the starting point for the rebuild, then trim the surfaces this project does not ship: Electron desktop packaging, the marketing website, and the ESP32 firmware tree. CI keeps the web build, e2e, lockfile, release, and Docker workflows. Rename the package and CLI entry points to hermes-ui, drop the prepare hook so npm ci no longer triggers a build, and drop the removed tsconfig and nodemon references. Rewrite AGENTS.md, ARCHITECTURE.md, DEVELOPMENT.md, and the READMEs so the documented package map and hard rules match this repository. reference_code/ is ignored: it is a read-only snapshot, not an input to the build.
63 lines
2.6 KiB
TypeScript
63 lines
2.6 KiB
TypeScript
import { describe, expect, it } from 'vitest'
|
|
import {
|
|
createMcuSpeechSegmenter,
|
|
normalizeMcuSpeechText,
|
|
} from '../../packages/server/src/services/global-agent/mcu-speech-segmenter'
|
|
|
|
describe('MCU speech segmenter', () => {
|
|
it('waits for markdown links to close before emitting speech', () => {
|
|
const segmenter = createMcuSpeechSegmenter({ maxChars: 24 })
|
|
|
|
expect(segmenter.pushDelta('请打开 [控制台')).toEqual([])
|
|
expect(segmenter.pushDelta('](https://example.com)。\n')).toEqual([
|
|
'请打开 控制台。',
|
|
])
|
|
})
|
|
|
|
it('skips fenced code and table rows when flushing readable text', () => {
|
|
const segmenter = createMcuSpeechSegmenter({ maxChars: 24 })
|
|
|
|
const segments = [
|
|
...segmenter.pushDelta('结果如下：\n```ts\nconst value = 1;\n```\n'),
|
|
...segmenter.pushDelta('| 名称 | 值 |\n| --- | --- |\n| foo | 1 |\n请确认。'),
|
|
]
|
|
const flushed = segmenter.flush()
|
|
expect(segments).toEqual([])
|
|
expect(flushed).toContain('结果如下')
|
|
expect(flushed).toContain('请确认')
|
|
expect(flushed).not.toContain('const value')
|
|
expect(flushed).not.toContain('foo')
|
|
})
|
|
|
|
it('emits completed paragraphs and carries unfinished paragraphs into the next one', () => {
|
|
const segmenter = createMcuSpeechSegmenter({ maxChars: 24 })
|
|
|
|
expect(segmenter.pushDelta('第一段还没结束\n')).toEqual([])
|
|
expect(segmenter.pushDelta('第二段结束了。\n')).toEqual([
|
|
'第一段还没结束 第二段结束了。',
|
|
])
|
|
expect(segmenter.pushDelta('第三段好了！\n')).toEqual([
|
|
'第三段好了！',
|
|
])
|
|
})
|
|
|
|
it('does not split long paragraphs on the old max character soft boundary', () => {
|
|
const segmenter = createMcuSpeechSegmenter({ maxChars: 24 })
|
|
const longText = '这是一段很长很长的内容，没有提前结束，也不会因为逗号，或者长度超过限制就提前播放'
|
|
|
|
expect(segmenter.pushDelta(longText)).toEqual([])
|
|
expect(segmenter.pushDelta('，直到段落正常结束。\n')).toEqual([
|
|
`${longText}，直到段落正常结束。`,
|
|
])
|
|
})
|
|
|
|
it('normalizes markdown without preserving table syntax', () => {
|
|
const normalized = normalizeMcuSpeechText('结果如下：\n| 名称 | 值 |\n| --- | --- |\n| foo | 1 |\n[详情](https://example.com)。参考 https://example.com/a?b=1 和 www.example.com/path')
|
|
expect(normalized).toContain('结果如下')
|
|
expect(normalized).toContain('详情')
|
|
expect(normalized).not.toContain('https')
|
|
expect(normalized).not.toContain('www.')
|
|
expect(normalized).not.toContain('foo')
|
|
})
|
|
})
|