CookGraph/start_docker.sh at main · QDcvd/CookGraph · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
#!/usr/bin/env bash
set -Eeuo pipefail

ROOT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
IMAGE_NAME="${IMAGE_NAME:-minicooking-agent-env}"
CONTAINER_NAME="${CONTAINER_NAME:-minicooking-agent}"
BACKEND_PORT="${BACKEND_PORT:-8000}"
FRONTEND_PORT="${FRONTEND_PORT:-5173}"
APT_MIRROR="${APT_MIRROR:-mirrors.aliyun.com}"
UV_INDEX_URL="${UV_INDEX_URL:-https://mirrors.aliyun.com/pypi/simple}"
NPM_REGISTRY="${NPM_REGISTRY:-https://registry.npmmirror.com}"
MODEL_SOURCE="${MODEL_SOURCE:-modelscope}"
MODELSCOPE_MODEL_ID="${MODELSCOPE_MODEL_ID:-AI-ModelScope/gte-large-zh}"
MODEL_REPO="${MODEL_REPO:-thenlper/gte-large-zh}"
HF_ENDPOINT="${HF_ENDPOINT:-https://hf-mirror.com}"
REBUILD="${REBUILD:-0}"

log() {
  printf '\033[1;32m[docker]\033[0m %s\n' "$*"
}

die() {
  printf '\033[1;31m[error]\033[0m %s\n' "$*" >&2
  exit 1
}

usage() {
  cat <<'EOF'
MiniCookingAgent-Demo Docker 一键启动

运行环境：
  仅支持 Ubuntu / WSL / Linux shell。
  请先进入 Ubuntu/WSL/Linux 环境，再在项目目录执行本脚本。

用法：
  bash start_docker.sh [选项]

选项：
  --rebuild          强制重新构建镜像
  --image NAME      镜像名，默认 minicooking-agent-env
  --backend-port N  宿主机后端端口，默认 8000
  --frontend-port N 宿主机前端端口，默认 5173
  -h, --help        显示帮助

常用环境变量：
  MODEL_SOURCE=modelscope|huggingface
  MODELSCOPE_MODEL_ID=AI-ModelScope/gte-large-zh
  MODEL_REPO=thenlper/gte-large-zh
  APT_MIRROR=mirrors.aliyun.com
  UV_INDEX_URL=https://mirrors.aliyun.com/pypi/simple
  NPM_REGISTRY=https://registry.npmmirror.com
EOF
}

case "$(uname -s 2>/dev/null || true)" in
  Linux*)
    ;;
  *)
    die "start_docker.sh 仅支持 Ubuntu / WSL / Linux shell。请先进入 Ubuntu/WSL/Linux 环境后再运行。"
    ;;
esac

while [[ $# -gt 0 ]]; do
  case "$1" in
    --rebuild)
      REBUILD=1
      shift
      ;;
    --image)
      IMAGE_NAME="${2:-}"
      [[ -n "$IMAGE_NAME" ]] || die "--image 需要镜像名"
      shift 2
      ;;
    --backend-port)
      BACKEND_PORT="${2:-}"
      [[ -n "$BACKEND_PORT" ]] || die "--backend-port 需要端口"
      shift 2
      ;;
    --frontend-port)
      FRONTEND_PORT="${2:-}"
      [[ -n "$FRONTEND_PORT" ]] || die "--frontend-port 需要端口"
      shift 2
      ;;
    -h|--help)
      usage
      exit 0
      ;;
    *)
      die "未知参数：$1"
      ;;
  esac
done

command -v docker >/dev/null 2>&1 || die "缺少 docker 命令，请先安装 Docker Desktop 或 Docker Engine。"

image_exists() {
  docker image inspect "$IMAGE_NAME" >/dev/null 2>&1
}

if [[ "$REBUILD" == "1" ]] || ! image_exists; then
  log "构建 Docker 镜像：$IMAGE_NAME"
  log "模型来源：$MODEL_SOURCE"
  docker build \
    --build-arg "APT_MIRROR=$APT_MIRROR" \
    --build-arg "UV_INDEX_URL=$UV_INDEX_URL" \
    --build-arg "NPM_REGISTRY=$NPM_REGISTRY" \
    --build-arg "MODEL_SOURCE=$MODEL_SOURCE" \
    --build-arg "MODELSCOPE_MODEL_ID=$MODELSCOPE_MODEL_ID" \
    --build-arg "MODEL_REPO=$MODEL_REPO" \
    --build-arg "HF_ENDPOINT=$HF_ENDPOINT" \
    -t "$IMAGE_NAME" \
    "$ROOT_DIR"
else
  log "复用已有镜像：$IMAGE_NAME（如需重建：bash start_docker.sh --rebuild）"
fi

log "启动容器，挂载项目：$ROOT_DIR -> /workspace"
log "访问地址：后端 http://localhost:$BACKEND_PORT，前端 http://localhost:$FRONTEND_PORT"

docker rm -f "$CONTAINER_NAME" >/dev/null 2>&1 || true

exec docker run --rm -it \
  --name "$CONTAINER_NAME" \
  -p "$BACKEND_PORT:8000" \
  -p "$FRONTEND_PORT:5173" \
  -v "$ROOT_DIR:/workspace" \
  -w /workspace \
  "$IMAGE_NAME" \
  python start.py --adapter agent_adapter_local_LLM_harness