Upload 2 files
Browse files- Dockerfile +2 -6
- sync_data.sh +93 -96
Dockerfile
CHANGED
|
@@ -21,7 +21,6 @@ ENV TZ=Asia/Shanghai
|
|
| 21 |
#ENV GID=1000
|
| 22 |
ENV ENABLE_AUTH=True
|
| 23 |
ENV WEBUI_AUTH=True
|
| 24 |
-
ENV ENABLE_SIGNUP=false
|
| 25 |
|
| 26 |
ENV DEFAULT_MODELS=gpt-4o-mini
|
| 27 |
ENV RAG_EMBEDDING_ENGINE=openai
|
|
@@ -42,15 +41,12 @@ ENV ENABLE_REALTIME_CHAT_SAVE=false
|
|
| 42 |
|
| 43 |
# 数据同步相关
|
| 44 |
ENV SYNC_INTERVAL=600
|
| 45 |
-
ENV DATASET_ID=deeme/ui
|
| 46 |
-
|
| 47 |
-
RUN apt-get update && apt-get install -y python3 python3-pip
|
| 48 |
-
RUN pip3 install --no-cache-dir huggingface_hub
|
| 49 |
|
| 50 |
COPY sync_data.sh sync_data.sh
|
| 51 |
|
| 52 |
RUN chmod -R 777 ./data && \
|
|
|
|
| 53 |
chmod -R 777 /app/backend/open_webui/static && \
|
| 54 |
chmod +x sync_data.sh && \
|
| 55 |
sed -i "1r sync_data.sh" ./start.sh && \
|
| 56 |
-
sed -i "1r redis-start.sh" ./start.sh
|
|
|
|
| 21 |
#ENV GID=1000
|
| 22 |
ENV ENABLE_AUTH=True
|
| 23 |
ENV WEBUI_AUTH=True
|
|
|
|
| 24 |
|
| 25 |
ENV DEFAULT_MODELS=gpt-4o-mini
|
| 26 |
ENV RAG_EMBEDDING_ENGINE=openai
|
|
|
|
| 41 |
|
| 42 |
# 数据同步相关
|
| 43 |
ENV SYNC_INTERVAL=600
|
|
|
|
|
|
|
|
|
|
|
|
|
| 44 |
|
| 45 |
COPY sync_data.sh sync_data.sh
|
| 46 |
|
| 47 |
RUN chmod -R 777 ./data && \
|
| 48 |
+
chmod -R 777 ./open_webui && \
|
| 49 |
chmod -R 777 /app/backend/open_webui/static && \
|
| 50 |
chmod +x sync_data.sh && \
|
| 51 |
sed -i "1r sync_data.sh" ./start.sh && \
|
| 52 |
+
sed -i "1r redis-start.sh" ./start.sh
|
sync_data.sh
CHANGED
|
@@ -1,113 +1,110 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
if [ -z "$HF_TOKEN" ] || [ -z "$DATASET_ID" ]; then
|
| 5 |
-
echo "缺少必要的环境变量 HF_TOKEN 或 DATASET_ID"
|
| 6 |
exit 1
|
| 7 |
fi
|
| 8 |
|
| 9 |
-
#
|
| 10 |
-
|
| 11 |
-
from huggingface_hub import HfApi
|
| 12 |
-
import sys
|
| 13 |
-
import os
|
| 14 |
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
api.delete_file(path_in_repo=file_to_delete, repo_id=repo_id, repo_type="dataset")
|
| 25 |
-
print(f'已删除旧备份: {file_to_delete}')
|
| 26 |
-
except Exception as e:
|
| 27 |
-
print(f'删除 {file_to_delete} 时出错: {str(e)}')
|
| 28 |
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
repo_id=repo_id,
|
| 36 |
-
repo_type="dataset"
|
| 37 |
-
)
|
| 38 |
-
print(f"成功上传 {file_name}")
|
| 39 |
-
|
| 40 |
-
manage_backups(api, repo_id)
|
| 41 |
-
except Exception as e:
|
| 42 |
-
print(f"文件上传出错: {str(e)}")
|
| 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 |
-
except Exception as e:
|
| 69 |
-
print(f"下载备份时出错: {str(e)}")
|
| 70 |
|
| 71 |
-
|
| 72 |
-
|
| 73 |
-
|
| 74 |
-
|
| 75 |
-
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
|
| 80 |
-
|
| 81 |
-
|
| 82 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 83 |
|
| 84 |
-
# 首次启动时下载最新备份
|
| 85 |
-
echo "正在从 HuggingFace 下载最新备份..."
|
| 86 |
-
python3 /tmp/hf_sync.py download "${HF_TOKEN}" "${DATASET_ID}"
|
| 87 |
|
| 88 |
-
# 同步函数
|
| 89 |
-
sync_data() {
|
| 90 |
-
while true; do
|
| 91 |
-
echo "开始同步进程 $(date)"
|
| 92 |
-
|
| 93 |
-
if [ -f "./data/webui.db" ]; then
|
| 94 |
-
timestamp=$(date +%Y%m%d_%H%M%S)
|
| 95 |
-
backup_file="webui_backup_${timestamp}.db"
|
| 96 |
-
|
| 97 |
-
# 复制数据库文件
|
| 98 |
-
cp ./data/webui.db "/tmp/${backup_file}"
|
| 99 |
-
|
| 100 |
-
echo "正在上传备份到 HuggingFace..."
|
| 101 |
-
python3 /tmp/hf_sync.py upload "${HF_TOKEN}" "${DATASET_ID}" "/tmp/${backup_file}" "${backup_file}"
|
| 102 |
-
|
| 103 |
-
rm -f "/tmp/${backup_file}"
|
| 104 |
else
|
| 105 |
-
|
|
|
|
|
|
|
|
|
|
| 106 |
fi
|
| 107 |
-
|
| 108 |
-
SYNC_INTERVAL=${SYNC_INTERVAL:-7200}
|
| 109 |
-
echo "
|
|
|
|
| 110 |
sleep $SYNC_INTERVAL
|
|
|
|
| 111 |
done
|
| 112 |
}
|
| 113 |
|
|
|
|
| 1 |
+
# 检查必要的环境变量
|
| 2 |
+
if [ -z "$G_NAME" ] || [ -z "$G_TOKEN" ]; then
|
| 3 |
+
echo "缺少必要的环境变量 G_NAME 或 G_TOKEN"
|
|
|
|
|
|
|
| 4 |
exit 1
|
| 5 |
fi
|
| 6 |
|
| 7 |
+
# 解析仓库名和用户名
|
| 8 |
+
IFS='/' read -r GITHUB_USER GITHUB_REPO <<< "$G_NAME"
|
|
|
|
|
|
|
|
|
|
| 9 |
|
| 10 |
+
# 构建 GitHub 仓库的克隆 URL,包含令牌
|
| 11 |
+
REPO_URL="https://${G_TOKEN}@github.com/${G_NAME}.git"
|
| 12 |
+
mkdir -p ./data/github_data
|
| 13 |
+
# 克隆仓库
|
| 14 |
+
echo "正在克隆仓库……"
|
| 15 |
+
git clone "$REPO_URL" ./data/github_data || {
|
| 16 |
+
echo "克隆失败,请检查 G_NAME 和 G_TOKEN 是否正确。"
|
| 17 |
+
exit 1
|
| 18 |
+
}
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
+
if [ -f ./data/github_data/webui.db ]; then
|
| 21 |
+
cp ./data/github_data/webui.db ./data/webui.db
|
| 22 |
+
echo "从 GitHub 仓库中拉取成功"
|
| 23 |
+
else
|
| 24 |
+
echo "GitHub 仓库中未找到 webui.db,将在同步时推送"
|
| 25 |
+
fi
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
|
| 27 |
+
# 定义同步函数
|
| 28 |
+
sync_data() {
|
| 29 |
+
while true; do
|
| 30 |
+
# 1. 同步到 GitHub
|
| 31 |
+
echo "正在开始同步"
|
| 32 |
+
# 进入仓库目录
|
| 33 |
+
cd ./data/github_data
|
| 34 |
+
# 配置 Git 用户信息
|
| 35 |
+
git config user.name "AutoSync Bot"
|
| 36 |
+
git config user.email "autosync@bot.com"
|
| 37 |
+
|
| 38 |
+
# 确保在正确的分支
|
| 39 |
+
git checkout main || git checkout master
|
| 40 |
+
|
| 41 |
+
# 复制最新的数据库文件
|
| 42 |
+
# cp ../webui.db ./webui.db
|
| 43 |
|
| 44 |
+
if [ -f "../webui.db" ]; then
|
| 45 |
+
cp ../webui.db ./webui.db
|
| 46 |
+
else
|
| 47 |
+
echo "数据库尚未初始化"
|
| 48 |
+
fi
|
| 49 |
|
| 50 |
+
# 检查是否有变化
|
| 51 |
+
if [[ -n $(git status -s) ]]; then
|
| 52 |
+
# 添加所有变更
|
| 53 |
+
git add webui.db
|
|
|
|
|
|
|
|
|
|
| 54 |
|
| 55 |
+
# 提交变更
|
| 56 |
+
git commit -m "Auto sync webui.db $(date '+%Y-%m-%d %H:%M:%S')"
|
| 57 |
+
|
| 58 |
+
# 推送到远程仓库
|
| 59 |
+
git push origin HEAD && {
|
| 60 |
+
echo "GitHub推送成功"
|
| 61 |
+
}|| {
|
| 62 |
+
echo "推送失败,等待重试..."
|
| 63 |
+
sleep 10
|
| 64 |
+
git push origin HEAD || {
|
| 65 |
+
echo "重试失败,放弃推送到Github。"
|
| 66 |
+
}
|
| 67 |
+
}
|
| 68 |
+
# 返回上级目录
|
| 69 |
+
cd ..
|
| 70 |
+
cd ..
|
| 71 |
+
|
| 72 |
+
# 2. 同步到 WebDAV
|
| 73 |
+
if [ -z "$WEBDAV_URL" ] || [ -z "$WEBDAV_USERNAME" ] || [ -z "$WEBDAV_PASSWORD" ]; then
|
| 74 |
+
echo "WebDAV 环境变量缺失,跳过 WebDAV 同步。"
|
| 75 |
+
else
|
| 76 |
+
echo "同步到 WebDAV..."
|
| 77 |
+
FILENAME="webui_$(date +'%m_%d').db"
|
| 78 |
+
# 检查是否存在要上传的文件
|
| 79 |
+
if [ -f ./data/webui.db ]; then
|
| 80 |
+
# 使用 curl 进行文件上传
|
| 81 |
+
curl -T ./data/webui.db --user "$WEBDAV_USERNAME:$WEBDAV_PASSWORD" "$WEBDAV_URL/$FILENAME" && {
|
| 82 |
+
echo "WebDAV 上传成功"
|
| 83 |
+
} || {
|
| 84 |
+
echo "WebDAV 上传失败,等待重试..."
|
| 85 |
+
sleep 10
|
| 86 |
+
curl -T ./data/webui.db --user "$WEBDAV_USERNAME:$WEBDAV_PASSWORD" "$WEBDAV_URL/$FILENAME" || {
|
| 87 |
+
echo "重试失败,放弃webdav上传。"
|
| 88 |
+
}
|
| 89 |
+
}
|
| 90 |
+
else
|
| 91 |
+
echo "未找到 webui.db 文件,跳过 WebDAV 同步"
|
| 92 |
+
fi
|
| 93 |
+
fi
|
| 94 |
|
|
|
|
|
|
|
|
|
|
| 95 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 96 |
else
|
| 97 |
+
# 返回上级目录
|
| 98 |
+
cd ..
|
| 99 |
+
cd ..
|
| 100 |
+
echo "GitHub: 没有检测到数据库变化"
|
| 101 |
fi
|
| 102 |
+
# 3. 等待统一的时间间隔
|
| 103 |
+
SYNC_INTERVAL=${SYNC_INTERVAL:-7200} # 默认间隔时间为 7200 秒
|
| 104 |
+
echo "当前时间 $(date '+%Y-%m-%d %H:%M:%S')"
|
| 105 |
+
echo "等待 ${SYNC_INTERVAL} 秒后进行下一次同步..."
|
| 106 |
sleep $SYNC_INTERVAL
|
| 107 |
+
|
| 108 |
done
|
| 109 |
}
|
| 110 |
|