Files
xiaoxia-saas/.env.example
T
xiaoxia 0cf7f14992
CI/CD Pipeline / Check push changed paths (pull_request) Has been skipped
CI/CD Pipeline / Build Staging API Image (pull_request) Has been skipped
CI/CD Pipeline / Dedup Check - skip PR tests when covered by push pipeline (pull_request) Successful in 2s
CI/CD Pipeline / Build Staging Web Image (pull_request) Has been skipped
CI/CD Pipeline / Check if frontend-only change (pull_request) Successful in 2s
CI/CD Pipeline / Build Staging Worker Image (pull_request) Has been skipped
CI/CD Pipeline / Retag skipped Staging API Image (pull_request) Has been skipped
CI/CD Pipeline / Retag skipped Staging Web Image (pull_request) Has been skipped
CI/CD Pipeline / Retag skipped Staging Worker Image (pull_request) Has been skipped
CI/CD Pipeline / Deploy Staging (Watchtower auto-deploy) (pull_request) Has been skipped
CI/CD Pipeline / Staging E2E Tests (pull_request) Has been skipped
CI/CD Pipeline / Staging API Integration Tests (pull_request) Has been skipped
CI/CD Pipeline / ACR Image Cleanup (pull_request) Has been skipped
CI/CD Pipeline / PR Build API Image (pull_request) Successful in 1m4s
PR Automation / Auto Approve on CI Green (pull_request) Successful in 3m17s
CI/CD Pipeline / PR Build Web Image (pull_request) Successful in 3m55s
CI/CD Pipeline / Validate - Python (mypy + alembic) (pull_request) Successful in 4m9s
CI/CD Pipeline / Frontend Lint (pull_request) Successful in 4m21s
CI/CD Pipeline / Build Production API Image (pull_request) Has been cancelled
CI/CD Pipeline / Build Production Web Image (pull_request) Has been cancelled
CI/CD Pipeline / Build Production Worker Image (pull_request) Has been cancelled
CI/CD Pipeline / Deploy Production (pull_request) Has been cancelled
CI/CD Pipeline / Production Browser E2E (pull_request) Has been cancelled
CI/CD Pipeline / Canary Release to Production (pull_request) Has been cancelled
CI/CD Pipeline / CI Gate (pull_request) Has been cancelled
CI/CD Pipeline / Validate - Style (pull_request) Has been cancelled
CI/CD Pipeline / Validate - Security (pull_request) Has been cancelled
CI/CD Pipeline / Unit Tests (pull_request) Has been cancelled
CI/CD Pipeline / Frontend Unit Tests (pull_request) Has been cancelled
CI/CD Pipeline / PR Build Worker Image (pull_request) Has been cancelled
CI/CD Pipeline / Integration Tests (pull_request) Has been cancelled
AI Code Review / AI Code Review (pull_request) Has been cancelled
PR Automation / Auto Merge on CI Green + Approved (pull_request) Has been cancelled
Preview Deploy / Deploy Preview Environment (pull_request) Has been cancelled
feat(viral-video): VLM提速+商品描述增强+爆款结构注入(问题4/5/6)
问题4:图片分析提速
- 新增 doubao_vision_lite_model 配置(默认 doubao-1-5-vision-lite-250915)+ doubao_vision_use_lite 开关
- call_vision 支持 model/max_tokens/temperature/timeout/system_prompt 参数覆盖
- _step_image_analysis 改 ThreadPoolExecutor 并行调用 VLM(max_workers=min(4,n_images)),lite 模型 timeout=30s,max_tokens=1024
- 图片直接传公网 OSS URL 给豆包视觉,无需本地下载-上传链路(原流程已是直传,保持)

问题5:VLM 商品描述增强(对齐竞品截图2粒度)
- 新 _IMAGE_ANALYSIS_SYSTEM_PROMPT 要求输出 16 个字段:name/brand/category/spec/appearance/packaging/text_on_package/colors/material_or_texture/key_features/visual_style/scene/suitable_scenes/target_audience/selling_points/summary
- appearance(外观80-150字)+packaging(包装80-150字)要求自然语言细节描述,summary(150-250字)直接给前端展示
- 严禁编造:看不清/没有的统一填「无法判断」,text_on_package 只 OCR 清晰可见的
- _build_products_summary 优先用 VLM 返回的 summary 段(自然段落给编导模型效果最好),兜底再拼结构化字段
- _analyze_single_image 独立函数,所有新字段都有 setdefault 防御,兼容旧字段名

问题6:爆款结构(copy_structure → viral_structure)透传
- 后端现有 schema 已支持 viral_structure 字段,STEP2 generate-copy 也已传,本次把 viral_structure 注入 _SCRIPT_GENERATION_PROMPT
- prompt 新增「爆款结构」字段+关键要求第8条:必须严格按该结构节奏/段落顺序编排镜头台词
- 前端 STRUCTURES 从 4 个扩展到 19 个(按协调 Agent 提供的中文列表)
- 一键路径 generateViralVideo 已传 viral_structure,无需改动

其它:
- .env.example 补 DOUBAO_VISION_MODEL/LITE_MODEL/USE_LITE 三个配置项
- scripts/render_env.sh 把新变量加入 SHARED_SECRETS 白名单

验收:单图分析目标 10-15s(lite 模型+并行),VLM 返回 6 维度丰富描述+summary,copy_structure 影响文案结构
2026-10-02 09:54:44 +08:00

259 lines
9.4 KiB
Bash
Executable File
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
# ============================================================
# 小虾 SaaS 环境变量完整配置
# ============================================================
# 本文件列出所有可配置的环境变量及默认值。
# 复制为 .env 后按需修改;生产环境务必覆盖所有密钥类配置。
#
# 配置读取规则(pydantic-settings,大小写不敏感):
# 1. 系统环境变量(最高优先级)
# 2. .env.{APP_ENV} 文件(如 .env.staging)
# 3. .env 文件
# 4. 代码中的默认值(最低优先级)
# ============================================================
# ==================== 应用基本配置 ====================
# 应用名称
APP_NAME=xiaoxia-saas
# 应用版本号(展示用,代码中已内置默认)
APP_VERSION=0.1.61
# 环境标识:development / staging / production
# 决定读取 .env.{APP_ENV} 还是 .env,也影响部分配置的严格校验
APP_ENV=development
# 是否开启 Debug 模式(开发环境 true,生产环境 false)
DEBUG=true
# 应用基础 URL,用于生成认证邮件、回调链接等
APP_BASE_URL=http://localhost:3000
# API 服务监听地址(容器内绑定,外部暴露由 Docker/Nginx 控制)
API_HOST=0.0.0.0
# API 服务监听端口
API_PORT=8000
# 是否自动创建数据库表结构(开发环境可开启,生产环境用 alembic migration)
AUTO_CREATE_SCHEMA=false
# ==================== 数据库配置 ====================
# 数据库连接串(格式:postgresql+psycopg://user:password@host:port/dbname)
DATABASE_URL=postgresql+psycopg://postgres:postgres@localhost:5432/xiaoxia_saas
# 连接池大小(常驻连接数)
DATABASE_POOL_SIZE=20
# 连接池最大溢出连接数(pool_size + max_overflow = 最大并发连接数)
DATABASE_MAX_OVERFLOW=10
# 获取连接超时时间(秒)
DATABASE_POOL_TIMEOUT=30
# 连接回收时间(秒),防止数据库端主动断开导致的死连接
DATABASE_POOL_RECYCLE=3600
# 是否使用内存数据库(SQLite,仅开发/测试可用;生产务必 false)
USE_IN_MEMORY_DB=false
# ==================== Redis 配置 ====================
# Redis 连接 URL(格式:redis://[:password@]host:port/db)
REDIS_URL=redis://localhost:6379/0
# 是否使用 Redis 存储 Session(多实例部署时必须开启;开发可用内存存储)
ENABLE_REDIS_SESSIONS=false
# ==================== Celery 任务队列 ====================
# Celery Broker(任务分发),默认用 Redis db0
CELERY_BROKER_URL=redis://localhost:6379/0
# Celery Result Backend(任务结果存储),默认用 Redis db1
CELERY_RESULT_BACKEND=redis://localhost:6379/1
# ==================== Worker 配置(#2073 队列分流) ====================
#
# 容器内跑三个独立进程:beat(只发定时任务)+ generation worker(实时高优)
# + transcode worker(后台批量/清理)。三个进程的并发与开关独立配置。
# Worker 进程名称
WORKER_NAME=xiaoxia-saas-worker
# 总并发参考(兼容旧变量):
# - 若 GENERATION_CONCURRENCY 与 TRANSCODE_CONCURRENCY 都未显式设置,
# entrypoint 会按此总数对半分配(gen=ceil(total/2), trans=剩余,各至少 1);
# - 任一个 *_CONCURRENCY 显式设置后,按显式值生效,忽略此变量对应部分。
WORKER_CONCURRENCY=4
# Generation worker 并发数(用户实时任务:视频生成/TTS/音色克隆/lipsync/数字人)
# 实时链路对延迟敏感,建议 2C 以上机器设为 2;高负载场景可加到 4。
GENERATION_CONCURRENCY=2
# Transcode worker 并发数(后台批量:素材入库转码/AI 分类打标/质量评分/查重/批量下载)
# 后台任务可排队,独立伸缩;素材入库量大时可加到 4。
TRANSCODE_CONCURRENCY=2
# 是否在本容器启动 celery beat 进程(默认 1)。
# 默认 beat 与 worker 同容器部署;若要独立 beat 容器部署,worker 容器设为 0、
# beat 容器单独跑 `celery -A worker_app.celery_app beat` 并设 BEAT_ENABLED=1。
BEAT_ENABLED=1
# 每个子进程最多处理多少任务后重启(防止内存泄漏)
WORKER_MAX_TASKS_PER_CHILD=1000
# ==================== JWT 认证配置 ====================
# JWT 签名密钥 — 生产环境必须设置为强随机字符串(至少32字符)
# 内置不安全值会被拒绝:secret / changeme / password / your-secret-key 等
JWT_SECRET_KEY=your-super-secret-key-change-this-in-production-min-32-chars
# JWT 签名算法
JWT_ALGORITHM=HS256
# Access Token 过期时间(分钟)
JWT_ACCESS_TOKEN_EXPIRE_MINUTES=30
# Refresh Token 过期时间(天)
JWT_REFRESH_TOKEN_EXPIRE_DAYS=30
# ==================== 邮件配置 ====================
# 是否启用邮件投递(关闭时邮件内容打印到日志,开发调试用)
ENABLE_EMAIL_DELIVERY=false
# SMTP 服务器地址
SMTP_HOST=smtp.gmail.com
# SMTP 端口
SMTP_PORT=587
# SMTP 用户名
SMTP_USER=your-email@gmail.com
# SMTP 密码 / 应用专用密码
SMTP_PASSWORD=your-app-specific-password
# 发件人邮箱
SMTP_FROM_EMAIL=noreply@xiaoxia-saas.com
# 发件人显示名称
SMTP_FROM_NAME=小虾 SaaS
# 是否启用 TLS
SMTP_USE_TLS=true
# ==================== 阿里云 OSS 配置 ====================
# OSS 区域 endpoint
OSS_ENDPOINT=oss-cn-hangzhou.aliyuncs.com
# OSS Access Key ID — 非开发环境必须设置
OSS_ACCESS_KEY_ID=your-access-key-id
# OSS Access Key Secret — 非开发环境必须设置
OSS_ACCESS_KEY_SECRET=your-access-key-secret
# OSS Bucket 名称
OSS_BUCKET_NAME=xiaoxia-autocut
# 直传最大文件大小(MB)
OSS_DIRECT_UPLOAD_MAX_MB=2000
# 直传签名有效期(秒)
OSS_DIRECT_UPLOAD_EXPIRE_SECONDS=900
# ==================== CORS 配置 ====================
# 允许跨域的前端域名列表,逗号分隔
CORS_ORIGINS_RAW=http://localhost:3000,http://localhost:5173,http://localhost:8000
# ==================== 渲染引擎配置 ====================
# 渲染引擎选择:
# legacy — 旧 VideoComposeService(稳定,功能完整)
# unified — 新 UnifiedRenderService(新架构,部分场景仍在验证)
RENDER_ENGINE=legacy
# ==================== CosyVoice 语音合成 ====================
# 阿里云百灵语音合成服务
# 模型选择:
# cosyvoice-v3-flash — 推荐,系统音色多,性价比高
# cosyvoice-v3-plus — 高质量,系统音色少
# cosyvoice-v3.5-flash / cosyvoice-v3.5-plus — 仅支持克隆/设计音色,无系统音色
# 音色:v3 系列系统音色带 _v3 后缀,如 longxiaochun_v3 / longxiaoxia_v3 / longanyang
COSYVOICE_API_KEY=your-cosyvoice-api-key
COSYVOICE_BASE_URL=https://dashscope.aliyuncs.com/api/v1
COSYVOICE_MODEL=cosyvoice-v3-flash
COSYVOICE_VOICE=longxiaochun_v3
COSYVOICE_SAMPLE_RATE=22050
COSYVOICE_FORMAT=mp3
# 音色克隆模型名(固定为 voice-enrollment,通常不需修改)
COSYVOICE_CLONE_MODEL=voice-enrollment
# ==================== 豆包大模型(火山引擎方舟) ====================
# 用于 AI 文案生成、智能剪辑等需要大模型能力的场景
DOUBAO_API_KEY=your-doubao-api-key
DOUBAO_MODEL=doubao-seed-1-6-250615
DOUBAO_BASE_URL=https://ark.cn-beijing.volces.com/api/v3
DOUBAO_TIMEOUT=30
DOUBAO_MAX_RETRIES=2
# 视觉模型:pro 精度高,lite 速度快(viral-video 商品识别默认用 lite 提速)
DOUBAO_VISION_MODEL=doubao-1-5-vision-pro-250915
DOUBAO_VISION_LITE_MODEL=doubao-1-5-vision-lite-250915
DOUBAO_VISION_USE_LITE=true
# ==================== 积分/会员系统 (#1895) ====================
# 积分系统总开关:默认 false(暂停积分系统)。
# - false:生成视频/口型同步/数字人/AI标题/TTS/克隆音色等所有功能对登录
# 用户免费放行,不扣积分、不做余额拦截;积分余额/流水/会员状态查询接口
# 保留可用,但数据不再变动。积分相关的表、代码、接口均保留不删除。
# - 恢复积分:设置 ENABLE_CREDIT_SYSTEM=true 即可,无需改代码。
ENABLE_CREDIT_SYSTEM=false
# 旧开关名(兼容别名):与 ENABLE_CREDIT_SYSTEM 任一为 true 即启用。
POINTS_ENABLED=false
# ==================== 抖音解析多源轮询 (#1963) ====================
# 无需配置 Key 也可使用(P0 免费源可用),配置 Key 可增加兜底能力
# TikHub API Key (https://tikhub.io) — $0.001/次起,注册送$0.05
TIKHUB_API_KEY=
# apizero.cn API Key (https://v1.apizero.cn) — 国内抖音解析服务
APIZERO_API_KEY=
# ==================== GPU MuseTalk Worker(反向轮询口型同步)====================
# GPU Worker 长期鉴权 Token,Worker 端 .env 的 GPU_WORKER_TOKEN 必须与此一致
# 留空时 development 环境允许匿名访问(仅本地调试),staging/production 必须配置
GPU_WORKER_TOKEN=
# 单任务超时(秒),processing 超过此时长无任务心跳才回退 pending 或标记 failed
# #1970:RTX2060 6G 推理 720p 长视频需 5 分钟以上,默认 900
GPU_TASK_TIMEOUT_SECONDS=900
# 是否启用 GPU 口型同步(开关)。开启后需同时有 Worker 在心跳窗口内(5分钟)才会走 GPU 路径;
# 开关关闭 / 无可用 Worker / GPU 任务失败或超时 → 自动回退现有 MediaKit 云端 lipsync
USE_GPU_LIPSYNC=false
# 业务侧轮询 GPU 任务结果的间隔(秒)
GPU_LIPSYNC_POLL_INTERVAL=5
# 业务侧等待 GPU 任务总超时(秒);超时回退 MediaKit
GPU_LIPSYNC_WAIT_TIMEOUT=1200
# Worker 心跳新鲜度窗口(秒),last_heartbeat_at 在此窗口内视为在线
GPU_WORKER_STALE_SECONDS=300