brucever commited on
Commit
2b7e3c5
·
verified ·
1 Parent(s): cdf5adc

Update Dockerfile

Browse files
Files changed (1) hide show
  1. Dockerfile +60 -70
Dockerfile CHANGED
@@ -1,93 +1,61 @@
1
  # 核心镜像:Node 22 slim 保证了环境的现代性与轻量化
2
  FROM node:22-slim
3
 
4
- # 1. 安装系统依赖
5
- # 包含:git (拉取依赖), openssh-client (解决构建报错), build-essential/g++/make (编译原生模块), python3 (运行同步脚本)
6
  RUN apt-get update && apt-get install -y --no-install-recommends \
7
- git openssh-client build-essential python3 python3-pip \
8
  g++ make ca-certificates \
9
  && rm -rf /var/lib/apt/lists/*
10
 
11
- # 2. 安装 Hugging Face 命令行工具
12
- RUN pip3 install --no-cache-dir huggingface_hub --break-system-packages
13
-
14
- # 3. 构建环境优化
15
- # 修复 Git 证书问题并将所有 SSH 协议重定向为 HTTPS
16
  RUN update-ca-certificates && \
17
  git config --global http.sslVerify false && \
18
  git config --global url."https://github.com/".insteadOf ssh://git@github.com/
19
 
20
- # 4. 全局安装 OpenClaw
21
  RUN npm install -g openclaw@latest --unsafe-perm
22
 
23
- # 5. 设置环境变量
24
  ENV PORT=7860 \
25
  OPENCLAW_GATEWAY_MODE=local \
26
- HOME=/root
 
27
 
28
- # 6. 核心同步引擎 (sync.py)
29
- # 针对 OpenClaw 新版 MEMORY.md 机制进行了全路径覆盖
30
- RUN echo 'import os, sys, tarfile\n\
31
- from huggingface_hub import HfApi, hf_hub_download\n\
32
- from datetime import datetime, timedelta\n\
33
- api = HfApi()\n\
34
- repo_id = os.getenv("HF_DATASET")\n\
35
- token = os.getenv("HF_TOKEN")\n\
36
- \n\
37
- def restore():\n\
38
- try:\n\
39
- print(f"--- [SYNC] 启动恢复流程, 目标仓库: {repo_id} ---")\n\
40
- if not repo_id or not token: \n\
41
- print("--- [SYNC] 跳过恢复: 未配置 HF_DATASET 或 HF_TOKEN ---")\n\
42
- return False\n\
43
- files = api.list_repo_files(repo_id=repo_id, repo_type="dataset", token=token)\n\
44
- now = datetime.now()\n\
45
- for i in range(5):\n\
46
- day = (now - timedelta(days=i)).strftime("%Y-%m-%d")\n\
47
- name = f"backup_{day}.tar.gz"\n\
48
- if name in files:\n\
49
- print(f"--- [SYNC] 发现备份文件: {name}, 正在下载... ---")\n\
50
- path = hf_hub_download(repo_id=repo_id, filename=name, repo_type="dataset", token=token)\n\
51
- with tarfile.open(path, "r:gz") as tar: tar.extractall(path="/root/.openclaw/")\n\
52
- print(f"--- [SYNC] 恢复成功! 数据已覆盖至 /root/.openclaw/ ---")\n\
53
- return True\n\
54
- print("--- [SYNC] 未找到最近 5 天的备份包 ---")\n\
55
- except Exception as e: print(f"--- [SYNC] 恢复异常: {e} ---")\n\
56
- \n\
57
- def backup():\n\
58
- try:\n\
59
- day = datetime.now().strftime("%Y-%m-%d")\n\
60
- name = f"backup_{day}.tar.gz"\n\
61
- print(f"--- [SYNC] 正在执行全量备份: {name} ---")\n\
62
- with tarfile.open(name, "w:gz") as tar:\n\
63
- # 路径说明:sessions(网关历史), workspace(记忆文件), agents(配置), memory(旧版目录)\n\
64
- for target in ["sessions", "workspace", "agents", "memory", "openclaw.json"]:\n\
65
- full_path = f"/root/.openclaw/{target}"\n\
66
- if os.path.exists(full_path):\n\
67
- tar.add(full_path, arcname=target)\n\
68
- api.upload_file(path_or_fileobj=name, path_in_repo=name, repo_id=repo_id, repo_type="dataset", token=token)\n\
69
- print(f"--- [SYNC] 备份上传成功! ---")\n\
70
- except Exception as e: print(f"--- [SYNC] 备份失败: {e} ---")\n\
71
- \n\
72
- if __name__ == "__main__":\n\
73
- if len(sys.argv) > 1 and sys.argv[1] == "backup": backup()\n\
74
- else: restore()' > /usr/local/bin/sync.py
75
 
76
- # 7. 容器入口脚本 (start-openclaw)
77
- # 负责恢复数据 -> 生成配置 -> 启动网关 -> 定时备份
78
  RUN echo "#!/bin/bash\n\
79
  set -e\n\
80
- mkdir -p /root/.openclaw/sessions\n\
81
- mkdir -p /root/.openclaw/workspace\n\
82
  \n\
83
- # 启动前执行数据恢复\n\
84
- python3 /usr/local/bin/sync.py restore\n\
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
85
  \n\
86
- # 清理 API Base 地址\n\
87
  CLEAN_BASE=\$(echo \"\$OPENAI_API_BASE\" | sed \"s|/chat/completions||g\" | sed \"s|/v1/|/v1|g\" | sed \"s|/v1\$|/v1|g\")\n\
88
  \n\
89
- # 生成 openclaw.json 配置文件\n\
90
- cat > /root/.openclaw/openclaw.json <<EOF\n\
 
91
  {\n\
92
  \"models\": {\n\
93
  \"providers\": {\n\
@@ -109,13 +77,35 @@ cat > /root/.openclaw/openclaw.json <<EOF\n\
109
  }\n\
110
  EOF\n\
111
  \n\
112
- # 启动定时备份进程 (每 3 小时执行一次,增强安全性)\n\
113
- (while true; do sleep 10800; python3 /usr/local/bin/sync.py backup; done) &\n\
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
114
  \n\
115
- # 启动 OpenClaw 网关\n\
 
116
  openclaw doctor --fix\n\
117
  exec openclaw gateway run --port \$PORT\n\
118
  " > /usr/local/bin/start-openclaw && chmod +x /usr/local/bin/start-openclaw
119
 
 
 
 
120
  EXPOSE 7860
121
  CMD ["/usr/local/bin/start-openclaw"]
 
1
  # 核心镜像:Node 22 slim 保证了环境的现代性与轻量化
2
  FROM node:22-slim
3
 
4
+ # 1. 安装系统依赖 (删除了 Python 和 pip,因为我们换成了 Git 备份)
 
5
  RUN apt-get update && apt-get install -y --no-install-recommends \
6
+ git openssh-client build-essential \
7
  g++ make ca-certificates \
8
  && rm -rf /var/lib/apt/lists/*
9
 
10
+ # 2. 构建环境优化 (保留原作者的证书修复)
 
 
 
 
11
  RUN update-ca-certificates && \
12
  git config --global http.sslVerify false && \
13
  git config --global url."https://github.com/".insteadOf ssh://git@github.com/
14
 
15
+ # 3. 全局安装 OpenClaw (保留原作者的命令)
16
  RUN npm install -g openclaw@latest --unsafe-perm
17
 
18
+ # 4. 环境与权限设置 (适配 Hugging Face 的普通用户规范)
19
  ENV PORT=7860 \
20
  OPENCLAW_GATEWAY_MODE=local \
21
+ HOME=/home/node \
22
+ OPENCLAW_HOME=/home/node/.openclaw
23
 
24
+ RUN mkdir -p /home/node/.openclaw/sessions \
25
+ && mkdir -p /home/node/.openclaw/workspace \
26
+ && mkdir -p /home/node/data \
27
+ && chown -R node:node /home/node
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
28
 
29
+ # 5. 容器入口脚本 (start-openclaw)
30
+ # 融合了我们的 Git 备份逻辑 + 原作者完美的 JSON 生成逻辑
31
  RUN echo "#!/bin/bash\n\
32
  set -e\n\
 
 
33
  \n\
34
+ # --- 1. 我们的 Git 数据恢复 ---\n\
35
+ if [ -n \"\$HF_TOKEN\" ] && [ -n \"\$DATASET_REPO\" ]; then\n\
36
+ echo \"🚀 初始化 Git 自动备份系统...\"\n\
37
+ git config --global user.email \"openclaw-bot@huggingface.co\"\n\
38
+ git config --global user.name \"OpenClaw Bot\"\n\
39
+ cd /home/node/data\n\
40
+ if [ ! -d \".git\" ]; then\n\
41
+ echo \"📥 克隆数据集: \$DATASET_REPO\"\n\
42
+ git clone https://user:\$HF_TOKEN@huggingface.co/datasets/\$DATASET_REPO .\n\
43
+ else\n\
44
+ echo \"📥 拉取最新数据...\"\n\
45
+ git pull origin main\n\
46
+ fi\n\
47
+ if [ \"\$(ls -A /home/node/data)\" ]; then\n\
48
+ echo \"♻️ 正在恢复历史数据...\"\n\
49
+ cp -r /home/node/data/* /home/node/.openclaw/ 2>/dev/null || true\n\
50
+ fi\n\
51
+ fi\n\
52
  \n\
53
+ # --- 2. 原作者的 API 清理逻辑 ---\n\
54
  CLEAN_BASE=\$(echo \"\$OPENAI_API_BASE\" | sed \"s|/chat/completions||g\" | sed \"s|/v1/|/v1|g\" | sed \"s|/v1\$|/v1|g\")\n\
55
  \n\
56
+ # --- 3. 原作者的 openclaw.json 生成逻辑 (一字不差保留) ---\n\
57
+ echo \"⚙️ 正在生成 OpenClaw 配置文件...\"\n\
58
+ cat > /home/node/.openclaw/openclaw.json <<EOF\n\
59
  {\n\
60
  \"models\": {\n\
61
  \"providers\": {\n\
 
77
  }\n\
78
  EOF\n\
79
  \n\
80
+ # --- 4. 解决局域网安全报错的魔法 ---\n\
81
+ if [ -n \"\$SPACE_HOST\" ]; then\n\
82
+ export OPENCLAW_GATEWAY_URL=\"wss://\$SPACE_HOST\"\n\
83
+ fi\n\
84
+ \n\
85
+ # --- 5. 我们的 Git 定时备份进程 (每5分钟后台执行) ---\n\
86
+ if [ -n \"\$HF_TOKEN\" ] && [ -n \"\$DATASET_REPO\" ]; then\n\
87
+ (while true; do \n\
88
+ sleep 300;\n\
89
+ cp -r /home/node/.openclaw/sessions /home/node/data/ 2>/dev/null || true;\n\
90
+ cp -r /home/node/.openclaw/workspace /home/node/data/ 2>/dev/null || true;\n\
91
+ cd /home/node/data;\n\
92
+ if [[ \$(git status --porcelain) ]]; then\n\
93
+ echo \"💾 [自动同步] 发现数据变动,正在上传...\"\n\
94
+ git add .;\n\
95
+ git commit -m \"Auto sync: \$(date)\";\n\
96
+ git push https://user:\$HF_TOKEN@huggingface.co/datasets/\$DATASET_REPO main > /dev/null 2>&1;\n\
97
+ fi;\n\
98
+ done) &\n\
99
+ fi\n\
100
  \n\
101
+ # --- 6. 原作者的启动逻辑 ---\n\
102
+ cd /home/node/.openclaw\n\
103
  openclaw doctor --fix\n\
104
  exec openclaw gateway run --port \$PORT\n\
105
  " > /usr/local/bin/start-openclaw && chmod +x /usr/local/bin/start-openclaw
106
 
107
+ # 6. 切换到普通用户运行 (必须项,否则 HF 会报错)
108
+ USER node
109
+ WORKDIR /home/node
110
  EXPOSE 7860
111
  CMD ["/usr/local/bin/start-openclaw"]