deepdata: delivery stack (docker compose, helm, CI, build scripts)

docker/docker-compose.yml

@@ -0,0 +1,533 @@
+services:
+  # ==============================
+  # Core DeepData Services (always enabled)
+  # ==============================
+
+  deepdata-backend:
+    container_name: deepdata-backend
+    #image: ${REGISTRY:-}deepdata-backend:${VERSION:-latest}
+    image: deepdata-backend:latest
+    restart: on-failure
+    privileged: true
+    environment:
+      - DB_PASSWORD=${DB_PASSWORD:?DB_PASSWORD is required. Set in .env file}
+      - JWT_SECRET=${JWT_SECRET:?JWT_SECRET is required. Set in .env file}
+      - deepdata.jwt.enable=${DEEPDATA_JWT_ENABLE:-false}
+      - LOG_ROTATION_MAX_SIZE=${LOG_ROTATION_MAX_SIZE:-100MB}
+      - LOG_ROTATION_BACKUP_COUNT=${LOG_ROTATION_BACKUP_COUNT:-30}
+    volumes:
+      - dataset_volume:/dataset
+      - flow_volume:/flow
+      - log_volume:/var/log/deepdata
+      - operator-upload-volume:/operators/upload
+      - operator-runtime-volume:/operators/extract
+    networks: [ deepdata ]
+    depends_on:
+      - deepdata-database
+
+  deepdata-backend-python:
+    container_name: deepdata-backend-python
+    #image: ${REGISTRY:-}deepdata-backend-python:${VERSION:-latest}
+    image: deepdata-backend-python:latest
+    restart: on-failure
+    privileged: true
+    ports:
+      - "18000:18000"
+    environment:
+      - log_level=DEBUG
+      - S3_READER_ALLOW_PRIVATE_IP=true
+      - DEEPDATA_SSRF_ALLOWED_HOSTS=172.16.249.172,minio,milvus-minio
+      - LOG_ROTATION_MAX_SIZE=${LOG_ROTATION_MAX_SIZE:-100MB}
+      - LOG_ROTATION_BACKUP_COUNT=${LOG_ROTATION_BACKUP_COUNT:-30}
+      - LOG_PVC_MONITOR_ENABLED=${LOG_PVC_MONITOR_ENABLED:-true}
+      - LOG_PVC_MONITOR_PATH=${LOG_PVC_MONITOR_PATH:-/var/log/deepdata}
+      - LOG_PVC_MONITOR_THRESHOLD=${LOG_PVC_MONITOR_THRESHOLD:-0.9}
+      - LOG_PVC_MONITOR_INTERVAL_SECONDS=${LOG_PVC_MONITOR_INTERVAL_SECONDS:-300}
+      - LOG_PVC_MONITOR_DELETE_BATCH_SIZE=${LOG_PVC_MONITOR_DELETE_BATCH_SIZE:-10}
+      - LOG_PVC_MONITOR_FILE_SUFFIXES=${LOG_PVC_MONITOR_FILE_SUFFIXES:-log,out,err}
+      - PGSQL_HOST=deepdata-database
+      - PGSQL_PORT=5432
+      - pgsql_password=${DB_PASSWORD:?DB_PASSWORD is required. Set in .env file}
+      - deepdata_jwt_enable=${DEEPDATA_JWT_ENABLE:-false}
+      - milvus_uri=${MILVUS_URI:-http://milvus:19530}
+      - LABEL_STUDIO_USERNAME=${LABEL_STUDIO_USERNAME:-admin@demo.com}
+      - LABEL_STUDIO_USER_TOKEN=${LABEL_STUDIO_USER_TOKEN:-}
+      - LABEL_STUDIO_PASSWORD=${LABEL_STUDIO_PASSWORD:-}
+      - LABEL_STUDIO_BASE_URL=${LABEL_STUDIO_BASE_URL:-http://label-studio:8000}
+      - COLLECTION_EXECUTION_MODE=${COLLECTION_EXECUTION_MODE:-local}
+    volumes:
+      #- ./template.py:/app/app/module/annotation/service/template.py:ro
+      #- ./client.py:/app/app/module/annotation/client/labelstudio/client.py:ro
+      #- ./project.py:/app/app/module/annotation/interface/project.py:ro
+      #- ./__init__.py:/app/app/module/rag/infra/embeddings/__init__.py:ro
+      #- ./store.py:/app/app/module/rag/infra/vectorstore/store.py:ro
+      #- ./factory.py:/app/app/module/shared/llm/factory.py:ro
+      #- ./file_processor.py:/app/app/module/rag/service/file_processor.py:ro
+      - dataset_volume:/dataset
+      - flow_volume:/flow
+      - log_volume:/var/log/deepdata
+      - graph_data_volume:/data/rag_storage
+      - operator-upload-volume:/operators/upload
+      - operator-runtime-volume:/operators/extract
+    networks: [ deepdata ]
+    depends_on:
+      #- deepdata-database
+      deepdata-database:
+        condition: service_healthy
+      milvus:
+        condition: service_healthy
+
+  deepdata-gateway:
+    container_name: deepdata-gateway
+    image: ${REGISTRY:-}deepdata-gateway:${VERSION:-latest}
+    restart: on-failure
+    privileged: true
+    ports:
+      - '8085:8080'
+    environment:
+      - JWT_SECRET=${JWT_SECRET:-}
+      - deepdata.jwt.enable=${DEEPDATA_JWT_ENABLE:-false}
+      - LOG_ROTATION_MAX_SIZE=${LOG_ROTATION_MAX_SIZE:-100MB}
+      - LOG_ROTATION_BACKUP_COUNT=${LOG_ROTATION_BACKUP_COUNT:-30}
+      - SPRING_DATASOURCE_URL=jdbc:postgresql://deepdata-database:5432/deepdata
+      - SPRING_DATASOURCE_USERNAME=postgres
+      - SPRING_DATASOURCE_PASSWORD=${DB_PASSWORD}
+      
+      # =====================================
+    volumes:
+      - log_volume:/var/log/deepdata
+    networks: [ deepdata ]
+
+  deepdata-frontend:
+    container_name: deepdata-frontend
+    image: ${REGISTRY:-}deepdata-frontend:${VERSION:-latest}
+    restart: on-failure
+    ports:
+      - "40000:3000"
+    environment:
+      - LOG_ROTATION_MAX_SIZE=${LOG_ROTATION_MAX_SIZE:-100MB}
+      - LOG_ROTATION_BACKUP_COUNT=${LOG_ROTATION_BACKUP_COUNT:-30}
+    volumes:
+      - ../frontend/dist:/opt/frontend/statics:ro
+      - frontend_log_volume:/var/log/deepdata/frontend
+    networks: [ deepdata ]
+    depends_on:
+      - deepdata-backend
+      - deepdata-backend-python
+
+  deepdata-database:
+    container_name: deepdata-database
+    image: ${REGISTRY:-}deepdata-database:${VERSION:-latest}
+    restart: on-failure
+    healthcheck:
+      test: ["CMD-SHELL", "pg_isready -U postgres"]
+      interval: 10s
+      timeout: 5s
+      retries: 5
+    environment:
+      - POSTGRES_USER=postgres
+      - POSTGRES_PASSWORD=${DB_PASSWORD:?DB_PASSWORD is required. Set in .env file}
+      - LOG_ROTATION_BACKUP_COUNT=${LOG_ROTATION_BACKUP_COUNT:-30}
+    volumes:
+      - postgresql_volume:/var/lib/postgresql/data
+      - database_log_volume:/var/log/deepdata/database
+    ports:
+      - "5442:5432"
+    networks: [ deepdata ]
+    
+  deepdata-runtime:
+    container_name: deepdata-runtime
+    image: deepdata-runtime-ray:latest
+    restart: on-failure
+    environment:
+      RAY_DEDUP_LOGS: "0"
+      RAY_TQDM_PATCH_PRINT: "0"
+      LOG_ROTATION_MAX_SIZE: ${LOG_ROTATION_MAX_SIZE:-100MB}
+      LOG_ROTATION_BACKUP_COUNT: ${LOG_ROTATION_BACKUP_COUNT:-30}
+      PG_HOST: "deepdata-database"
+      PG_PORT: "5432"
+      PG_USER: "postgres"
+      PG_PASSWORD: ${DB_PASSWORD:?DB_PASSWORD is required. Set in .env file}
+      PG_DATABASE: "deepdata"
+      RAY_DASHBOARD_ADDRESS: ${RAY_DASHBOARD_ADDRESS}
+      RAY_REDIS_PASSWORD: ${RAY_REDIS_PASSWORD}
+      RAY_REDIS_ADDRESS: ${RAY_REDIS_ADDRESS}
+      ENABLE_RAY_EXECUTION: ${ENABLE_RAY_EXECUTION:-false}
+    command:
+      - python
+      - /opt/runtime/deepdata/operator_runtime.py
+      - --port
+      - "8081"
+    volumes:
+      - ray_log_volume:/tmp/ray
+      - log_volume:/var/log/deepdata
+      - dataset_volume:/dataset
+      - flow_volume:/flow
+      - operator-runtime-volume:/opt/runtime/deepdata/ops/user
+      - operator-packages-volume:/usr/local/lib/ops/site-packages
+    networks: [ deepdata ]
+
+  # =============================
+  # Optional: Mineru NPU Engine (profile: mineru)
+  # ==============================
+  deepdata-mineru:
+    container_name: deepdata-mineru
+    image: deepdata-mineru
+    restart: on-failure
+    environment:
+      MINERU_MODEL_SOURCE: local
+      MINERU_DEVICE_MODE: npu
+      VLLM_WORKER_MULTIPROC_METHOD: spawn
+    privileged: true
+    entrypoint: mineru-openai-server
+    command:
+      --engine vllm
+      --host 0.0.0.0
+      --port 8000
+      ${EXTRA_ARGS:-}
+    volumes:
+      - /var/log/npu/:/usr/slog
+      - /usr/local/dcmi:/usr/local/dcmi
+      - /usr/local/bin/npu-smi:/usr/local/bin/npu-smi
+      - /usr/local/Ascend/driver:/usr/local/Ascend/driver
+    networks: [ deepdata ]
+    profiles: [ mineru ]
+    devices:
+      - /dev/davinci0
+      - /dev/davinci_manager
+      - /dev/devmm_svm
+      - /dev/hisi_hdc
+
+  # ==============================
+  # Optional: Redis (profile: redis)
+  # ==============================
+  deepdata-redis:
+    container_name: deepdata-redis
+    image: redis:8.2.3
+    restart: on-failure
+    ports:
+      - "6390:6379"
+    networks: [ deepdata ]
+    profiles: [ redis ]
+
+  # ==============================
+  # Optional: ray
+  # ==============================
+
+  deepdata-ray-redis:
+    container_name: deepdata-ray-redis
+    image: redis:8.2.3
+    restart: always
+    command: redis-server --requirepass ${RAY_REDIS_PASSWORD}
+    ports:
+      - "6399:6379"
+    networks:
+      - deepdata
+    volumes:
+      - ray_redis_data:/data
+    profiles: [ ray ]
+
+  deepdata-ray-head:
+    container_name: deepdata-ray-head
+    image: deepdata-runtime-ray:latest
+    restart: always
+    ports:
+      - "${RAY_DASHBOARD_PORT}:8265"    # Dashboard UI
+      - "${RAY_CLIENT_PORT}:10001"      # Client Submission
+    environment:
+      - RAY_REDIS_PASSWORD=${RAY_REDIS_PASSWORD}
+      - RAY_REDIS_ADDRESS=${RAY_REDIS_ADDRESS}
+      - RAY_MODE=head
+    networks:
+      - deepdata
+    shm_size: '2gb'
+    volumes:
+      - dataset_volume:/dataset
+      - flow_volume:/flow
+      - ray_log_volume:/tmp/ray
+    depends_on:
+      - deepdata-ray-redis
+    profiles: [ ray ]
+
+  deepdata-ray-worker:
+    image: deepdata-runtime-ray:latest
+    restart: always
+    deploy:
+      replicas: 2
+    shm_size: '8gb'
+    environment:
+      - RAY_MODE=worker
+      - RAY_HEAD_ADDRESS=${RAY_HEAD_GCS_ADDRESS} 
+      - RAY_REDIS_PASSWORD=${RAY_REDIS_PASSWORD}
+      - NUM_CPUS=${RAY_WORKER_NUM_CPUS:-4}
+      - LOG_ROTATION_MAX_SIZE=${LOG_ROTATION_MAX_SIZE:-100MB}
+      - LOG_ROTATION_BACKUP_COUNT=${LOG_ROTATION_BACKUP_COUNT:-30}
+      - PG_HOST=deepdata-database
+      - PG_PORT=5432
+      - PG_USER=postgres
+      - PG_PASSWORD=${DB_PASSWORD:?DB_PASSWORD is required. Set in .env file}
+      - PG_DATABASE=deepdata
+    volumes:
+      - dataset_volume:/dataset
+      - flow_volume:/flow
+      - ray_log_volume:/tmp/ray
+    networks: [ deepdata ]
+    depends_on:
+      - deepdata-ray-head
+    profiles: [ray]
+
+  # ==============================
+  # Optional: Data Juicer (profile: data-juicer)
+  # ==============================
+  deepdata-data-juicer:
+    container_name: deepdata-data-juicer
+    image: datajuicer/data-juicer:v1.4.4
+    restart: on-failure
+    command:
+      - uvicorn
+      - service:app
+      - --host
+      - "0.0.0.0"
+    volumes:
+      - dataset_volume:/dataset
+      - flow_volume:/flow
+    networks: [ deepdata ]
+    profiles: [ data-juicer ]
+
+
+  # ==============================
+  # Optional: Deer Flow (profile: deepflow)
+  # ==============================
+  deepflow-backend:
+    # image: ${REGISTRY:-}deepflow-backend
+    image: deepflow-backend:latest
+    container_name: deepflow-backend
+    env_file:
+      - ./deepflow/.env
+    volumes:
+      - ./deepflow/conf.yaml:/app/conf.yaml:ro
+      - deepflow-log-volume:/var/log/deepflow
+      - dataset_volume:/dataset
+    restart: on-failure
+    networks:
+      - deepdata
+    profiles:
+      - deepflow
+
+  deepflow-frontend:
+    #image: ${REGISTRY:-}deepflow-frontend
+    image: deepflow-frontend:latest
+    container_name: deepflow-frontend
+    env_file:
+      - ./deepflow/.env
+    depends_on:
+      - deepflow-backend
+    restart: on-failure
+    networks:
+      - deepdata
+    profiles:
+      - deepflow
+
+  # ==============================
+  # Optional: Label Studio (profile: label-studio)
+  # ==============================
+  label-studio-pgbouncer:
+    container_name: label-studio-pgbouncer
+    image: edoburu/pgbouncer:latest
+    restart: on-failure
+    environment:
+      - DB_HOST=deepdata-database
+      - DB_PORT=5432
+      - DB_NAME=labelstudio
+      - DB_USER=postgres
+      - DB_PASSWORD=${DB_PASSWORD:?DB_PASSWORD is required. Set in .env file}
+      - AUTH_TYPE=scram-sha-256
+      - POOL_MODE=transaction
+      - MAX_CLIENT_CONN=100
+      - DEFAULT_POOL_SIZE=20
+      - MAX_DB_CONNECTIONS=20
+    networks:
+      - deepdata
+    profiles: [ label-studio ]
+
+  label-studio:
+    container_name: label-studio
+    stdin_open: true
+    tty: true
+    image: heartexlabs/label-studio:latest
+    privileged: true
+    restart: on-failure
+    user: root
+    expose:
+      - "8000"
+    # M4 入口收敛：删除 host 直连端口 38000:8000（原 L364），仅保留内部网络端口，
+    # 外部访问一律走网关（详见 integration/deployment/label-studio入口收敛.md）
+    depends_on:
+      - label-studio-pgbouncer
+    environment:
+      - DJANGO_DB=default
+      - POSTGRE_NAME=labelstudio
+      - POSTGRE_USER=postgres
+      - POSTGRE_PASSWORD=${DB_PASSWORD:?DB_PASSWORD is required. Set in .env file}
+      - POSTGRE_PORT=5432
+      - POSTGRE_HOST=label-studio-pgbouncer
+      - LABEL_STUDIO_HOST=${LABEL_STUDIO_HOST:-}
+      - LOCAL_FILES_SERVING_ENABLED=true
+      - LOCAL_FILES_DOCUMENT_ROOT=/label-studio/local
+      - USE_USERNAME_FOR_LOGIN=true
+      - LABEL_STUDIO_USERNAME=admin@demo.com
+      - LABEL_STUDIO_PASSWORD=${LABEL_STUDIO_PASSWORD:-}
+      - LABEL_STUDIO_ENABLE_LEGACY_API_TOKEN=true
+      - LABEL_STUDIO_USER_TOKEN=${LABEL_STUDIO_USER_TOKEN:-}
+      - LOG_LEVEL=DEBUG
+    volumes:
+      - label-studio-data:/label-studio/data:rw
+      - dataset_volume:/label-studio/local:rw
+    networks:
+      - deepdata
+    command: label-studio-uwsgi
+    profiles: [ label-studio ]
+
+  # ==============================
+  # Optional: Milvus (profile: milvus)
+  # ==============================
+  etcd:
+    container_name: milvus-etcd
+    #image: quay.io/coreos/etcd:v3.5.18
+    image: quay.io/coreos/etcd:v3.5.9
+    environment:
+      - ETCD_AUTO_COMPACTION_MODE=revision
+      - ETCD_AUTO_COMPACTION_RETENTION=1000
+      - ETCD_QUOTA_BACKEND_BYTES=4294967296
+      - ETCD_SNAPSHOT_COUNT=50000
+    volumes:
+      - etcd-volume:/etcd
+    command: etcd -advertise-client-urls=http://etcd:2379 -listen-client-urls http://0.0.0.0:2379 --data-dir /etcd
+    restart: always
+    networks:
+      - deepdata
+    healthcheck:
+      test: ["CMD", "etcdctl", "endpoint", "health"]
+      interval: 30s
+      timeout: 20s
+      retries: 3
+    profiles:
+      - milvus
+
+  minio:
+    container_name: milvus-minio
+    image: minio/minio:RELEASE.2024-12-18T13-15-44Z
+    environment:
+      MINIO_ACCESS_KEY: ${MINIO_ACCESS_KEY:-}
+      MINIO_SECRET_KEY: ${MINIO_SECRET_KEY:-}
+    ports:
+      - "9001:9001"
+      - "9000:9000"
+    volumes:
+      - minio-volume:/data
+    command: minio server /data --console-address ":9001"
+    networks:
+      - deepdata
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
+      interval: 30s
+      timeout: 20s
+      retries: 3
+    profiles:
+      - milvus
+
+  milvus:
+    container_name: milvus-standalone
+    #restart: on-failure
+    restart: unless-stopped
+    image: milvusdb/milvus:v2.6.5
+    command: ["milvus", "run", "standalone"]
+    security_opt:
+    - seccomp:unconfined
+    environment:
+      ETCD_ENDPOINTS: etcd:2379
+      MINIO_ADDRESS: minio:9000
+      MQ_TYPE: woodpecker
+      MINIO_ACCESS_KEY_ID: ${MINIO_ACCESS_KEY_ID}
+      MINIO_SECRET_ACCESS_KEY: ${MINIO_SECRET_ACCESS_KEY}
+      MINIO_USE_SSL: "false"
+      AWS_REGION: ${AWS_REGION}
+    volumes:
+      - milvus-volume:/var/lib/milvus
+    networks:
+      - deepdata
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:9091/healthz"]
+      interval: 30s
+      start_period: 180s
+      #start_period: 90s
+      timeout: 20s
+      retries: 10
+      # retries: 3
+    ports:
+      - "19530:19530"
+      - "19091:9091"
+    depends_on:
+      #- "etcd"
+      #- "minio"
+      etcd:
+        condition: service_healthy   # ✅ 必须
+      minio:
+        condition: service_healthy   # ✅ 必须
+    profiles:
+      - milvus
+
+# ==============================
+# Volumes
+# ==============================
+volumes:
+  dataset_volume:
+    name: deepdata-dataset-volume
+  flow_volume:
+    name: deepdata-flow-volume
+  log_volume:
+    name: deepdata-log-volume
+  postgresql_volume:
+    name: deepdata-postgresql-volume
+  ray_log_volume:
+    name: deepdata-ray-log-volume
+  frontend_log_volume:
+    name: deepdata-frontend-log-volume
+  database_log_volume:
+    name: deepdata-database-log-volume
+  operator-upload-volume:
+    name: deepdata-operator-upload-volume
+  operator-runtime-volume:
+    name: deepdata-operator-runtime-volume
+  operator-packages-volume:
+    name: deepdata-operator-packages-volume
+  graph_data_volume:
+    name: deepdata-graph-data-volume
+
+  # Deer Flow
+  deepflow-log-volume:
+    name: deepflow-log-volume
+
+  # Label Studio
+  label-studio-data:
+    name: label-studio-data-volume
+  label-studio-db:
+    name: label-studio-db-volume
+
+  # Milvus
+  etcd-volume:
+    name: milvus-etcd-volume
+  minio-volume:
+    name: milvus-minio-volume
+  milvus-volume:
+    name: milvus-milvus-volume
+  ray_redis_data:
+    name: deepdata-ray-redis-data
+
+# ==============================
+# Networks
+# ==============================
+networks:
+  deepdata:
+    driver: bridge
+    name: deepdata-network