 灵应
|
2849123bd7
|
feat: P2 WebSocket 流式合成
CI/CD Pipeline / Deploy Staging (push) Has been skipped
CI/CD Pipeline / Staging E2E Tests (push) Has been skipped
CI/CD Pipeline / Build Production Runtime Images (push) Has been skipped
CI/CD Pipeline / Deploy Production (push) Has been skipped
CI/CD Pipeline / Production Browser E2E (push) Has been skipped
CI/CD Pipeline / Frontend Lint (push) Failing after 47h49m26s
CI/CD Pipeline / Validate Code Quality And Tests (push) Failing after 47h49m26s
- 新增 TTSStreamingService 流式合成服务
- 实现 /ws/tts/stream WebSocket 端点
- 短文本（≤500字）：合成完整音频后分块推送
- 长文本（>500字）：分段并发合成，逐段推送
- 通过 asyncio.to_thread 桥接同步 CosyVoiceService
- 二进制帧 4KB 分块推送音频数据
- 10 个单元测试全部通过
|
2026-07-07 13:04:59 +08:00 |
|
|
灵应
|
d6ab413dcd
|
feat: P1 长文本分段合成
CI/CD Pipeline / Deploy Staging (push) Has been skipped
CI/CD Pipeline / Staging E2E Tests (push) Has been skipped
CI/CD Pipeline / Build Production Runtime Images (push) Has been skipped
CI/CD Pipeline / Deploy Production (push) Has been skipped
CI/CD Pipeline / Production Browser E2E (push) Has been skipped
CI/CD Pipeline / Frontend Lint (push) Failing after 47h57m37s
CI/CD Pipeline / Validate Code Quality And Tests (push) Failing after 47h57m37s
- 新增 text_splitter.py: 长文本智能分段（句子边界 + 短段合并）
- 新增 audio_merger.py: FFmpeg concat 音频合并器
- workflow.py: 分段合成完整流程（同步合并 / 异步轮询 / 失败处理）
- tts_synthesis.py: 新增 process_tts_segment_synthesis Celery 任务
- tts.py: 路由层自动识别分段任务并分发到对应 Celery task
- 23 个单元测试全部通过，P0 回归测试无退化
|
2026-07-07 12:56:52 +08:00 |
|
|
灵应
|
c47aed95fc
|
feat(tts): P0 音频转存 OSS — 合成完成后下载临时音频上传 OSS 存储永久 URL
CI/CD Pipeline / Deploy Staging (push) Has been skipped
CI/CD Pipeline / Staging E2E Tests (push) Has been skipped
CI/CD Pipeline / Build Production Runtime Images (push) Has been skipped
CI/CD Pipeline / Deploy Production (push) Has been skipped
CI/CD Pipeline / Production Browser E2E (push) Has been skipped
CI/CD Pipeline / Frontend Lint (push) Failing after 48h27m21s
CI/CD Pipeline / Validate Code Quality And Tests (push) Failing after 48h27m21s
- TTSWorkflowService 新增 _transfer_audio_to_oss 方法
- start_synthesis 同步路径和 process_synthesis_result 均集成 OSS 转存
- OSS 转存失败时优雅降级，回退到 CosyVoice 临时 URL
- 支持 mp3/wav/pcm/opus 格式 content-type 映射
- 新增 9 个单元测试覆盖转存成功/失败/降级场景
- 向后兼容：storage_service 为可选参数，现有调用方无需修改
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
|
2026-07-07 12:27:06 +08:00 |
|
|
灵应
|
8935196fcd
|
style: 后端代码black格式化
Deploy / Staging E2E Tests (push) Has been skipped
Deploy / Build Production Runtime Images (push) Has been skipped
Deploy / Deploy Production (push) Has been skipped
Deploy / Production Browser E2E (push) Has been skipped
Deploy / Deploy Staging (push) Failing after 138h4m33s
CI/CD Pipeline / Frontend Lint (push) Failing after 138h4m39s
CI/CD Pipeline / Validate Code Quality And Tests (push) Failing after 138h4m39s
|
2026-07-03 18:49:54 +08:00 |
|
|
灵应
|
e539105256
|
fix: 3.09 审计问题修复 — P1-1 TTS Celery 执行器 + P2-1~P2-4
Deploy / Build Production Runtime Images (push) Has been skipped
Deploy / Deploy Production (push) Has been skipped
Deploy / Production Browser E2E (push) Has been skipped
CI/CD Pipeline / Frontend Lint (pull_request) Failing after 163h53m1s
CI/CD Pipeline / Validate Code Quality And Tests (pull_request) Failing after 163h53m6s
Deploy / Deploy Staging (push) Failing after 164h23m43s
CI/CD Pipeline / Frontend Lint (push) Failing after 164h23m43s
CI/CD Pipeline / Validate Code Quality And Tests (push) Failing after 164h23m50s
P1-1 (阻塞性): TTS 合成完整异步链路
- CosyVoiceService 新增 submit_synthesize_task() + poll_synthesize_task()
- 新建 TTSWorkflowService 编排层 (packages/application/tts_job/workflow.py)
- 新建 Celery 任务 process_tts_synthesis (apps/worker/worker_app/tasks/tts_synthesis.py)
- 注册到 celery_app.conf.imports + tasks/__init__.py 懒加载
- TTS 路由 synthesize() 增加 CosyVoice 提交 + Celery 调度
P2-1: voice_clone.py 添加详细 Celery 重试策略注释
P2-2: 修复 voice_clone.py Session 泄漏 (session=None 安全模式)
P2-3: ListVoiceLibraryUseCase 返回 (items, count) 元组，消除重复 count_by_user()
P2-4: 新增 find_profile_ids_by_voice_ids() 批量方法，填充 voice_clone_profile_id
测试: 749 passed, 0 failed
|
2026-07-02 16:30:37 +08:00 |
|
|
灵应
|
e29a698325
|
feat: implement TTS synthesis API (task 3.06)
Deploy / Build Production Runtime Images (push) Has been skipped
Deploy / Deploy Production (push) Has been skipped
Deploy / Production Browser E2E (push) Has been skipped
CI/CD Pipeline / Frontend Lint (pull_request) Failing after 169h43m28s
CI/CD Pipeline / Validate Code Quality And Tests (pull_request) Failing after 169h43m33s
Deploy / Deploy Staging (push) Failing after 169h47m22s
CI/CD Pipeline / Frontend Lint (push) Failing after 169h47m58s
CI/CD Pipeline / Validate Code Quality And Tests (push) Failing after 169h48m6s
- POST /api/v1/tts/synthesize: create TTS job (status=pending)
- GET /api/v1/tts/jobs: list user's jobs (page/page_size pagination)
- GET /api/v1/tts/jobs/{job_id}: get job details
- GET /api/v1/tts/jobs/{job_id}/status: query status for polling
- DELETE /api/v1/tts/jobs/{job_id}: delete job (soft delete)
Includes:
- Schema: TTSSynthesizeRequest, TTSJobResponse, ListTTSJobResponse, TTSStatusResponse
- Use Cases: CreateTTSJob, ListTTSJobs, GetTTSJob, GetTTSJobStatus, DeleteTTSJob
- SQLAlchemy adapter: SQLAlchemyTTSJobRepository
- ORM model: TTSJobModel
- Alembic migration: 020_add_tts_jobs_table
- Unit tests: 14 tests covering all use cases
CosyVoice API calls deferred to task 3.07.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
|
2026-07-02 11:06:25 +08:00 |
|