自建 SearXNG 私有聚合搜索引擎：Docker 部署与无广告搜索

# ================= 全局参数定义 =================
DOMAIN="search.0000996.xyz"     # 访问 SearXNG 的专属公网域名
DATA_DIR="/opt/searxng"         # 持久化配置与数据根目录
PORT="8080"                     # SearXNG 容器绑定的本地回环端口
# ================================================

CRITICAL:searx.configuration: secret_key is not set in settings.yml
Error: Exiting due to missing secret_key configuration.

sudo mkdir -p /opt/searxng/searxng
cd /opt/searxng

# 生成高强度随机密钥
openssl rand -hex 32

sudo tee /opt/searxng/searxng/settings.yml << 'EOF'
# --- SearXNG 核心基础设置 ---
use_default_settings: true

general:
  debug: false
  instance_name: "Private Metasearch"
  donation_url: false
  contact_url: false
  privacypolicy_url: false
  enable_metrics: false

# --- Web 前端与主题配置 ---
ui:
  static_use_hash: true
  default_locale: "zh-CN"
  query_in_title: true
  infinite_scroll: true          # 开启无缝瀑布流翻页
  center_alignment: true         # 搜索框与结果居中排版 (极客美学)
  default_theme: "simple"
  theme_args:
    simple_style: "dark"         # 默认开启护眼暗黑极客主题
  search_on_category_select: true
  hotkeys: "vim"                 # 支持 Vim 快捷键导航 (j/k 翻选)

# --- 搜索行为与请求调优 ---
search:
  safe_search: 0                 # 0: 关闭安全过滤, 1: 中度, 2: 严格
  autocomplete: "google"         # 搜索词自动补全源 (支持 google, bing, duckduckgo)
  ban_time_on_fail: 5
  max_ban_time_on_fail: 120
  formats:
    - html
    - json                       # 核心：必须解禁 json 格式，否则 AI Agent 调用报 403

# --- 生产服务监听与安全密钥 ---
server:
  port: 8080
  bind_address: "0.0.0.0"
  secret_key: "9f8a3c2b1e4d5a6b7c8d9e0f1a2b3c4d5e6f7a8b9c0d1e2f3a4b5c6d7e8f9a0b" # 替换为你自己的密钥
  base_url: "https://search.0000996.xyz/" # 你的实际公网反代根地址
  limiter: false                 # 私有实例建议关闭限流器，避免 AI Agent 频繁请求被关小黑屋
  image_proxy: true              # 开启图片代理，防止加载外部图片泄露本地真实 IP
  method: "POST"                 # 优先使用 POST 发送查询，增强私密性

# --- 常用主力引擎权重调优 ---
engines:
  - name: google
    engine: google
    shortcut: g
    weight: 2.0
    timeout: 3.0

  - name: bing
    engine: bing
    shortcut: b
    weight: 1.5
    timeout: 3.0

  - name: duckduckgo
    engine: duckduckgo
    shortcut: d
    weight: 1.2
    timeout: 3.0

  - name: github
    engine: github
    shortcut: gh
    weight: 1.0
    categories: it

  - name: wikidata
    engine: wikidata
    shortcut: w
    weight: 1.0
EOF

sudo tee /opt/searxng/docker-compose.yml << 'EOF'
version: '3.8'

services:
  # --- 本地高速缓存服务 (Redis) ---
  redis:
    image: redis:alpine
    container_name: searxng-redis
    restart: always
    command: redis-server --save "" --appendonly no  # 仅作为瞬态内存缓存，关闭持久化保护硬盘
    tmpfs:
      - /var/run/redis:rw,noexec,nosuid,size=64m
    logging:
      driver: "json-file"
      options:
        max-size: "10m"
        max-file: "2"

  # --- SearXNG 主引擎容器 ---
  searxng:
    image: searxng/searxng:latest
    container_name: searxng
    restart: always
    depends_on:
      - redis
    ports:
      - "127.0.0.1:8080:8080"  # 仅绑定本地回环，安全交由反向代理转发
    volumes:
      - ./searxng:/etc/searxng:rw
    environment:
      - SEARXNG_BASE_URL=https://search.0000996.xyz/
    cap_drop:
      - ALL
    cap_add:
      - CHOWN
      - SETGID
      - SETUID
    logging:
      driver: "json-file"
      options:
        max-size: "20m"
        max-file: "3"
EOF

cd /opt/searxng
sudo docker compose up -d

sudo docker compose ps

server {
    listen 80;
    server_name search.0000996.xyz;
    return 301 https://$host$request_uri;
}

server {
    listen 443 ssl http2;
    server_name search.0000996.xyz;

    # SSL 证书配置路径
    ssl_certificate     /etc/nginx/ssl/search.0000996.xyz.crt;
    ssl_certificate_key /etc/nginx/ssl/search.0000996.xyz.key;
    ssl_protocols       TLSv1.2 TLSv1.3;
    ssl_ciphers         HIGH:!aNULL:!MD5;

    # 隐私防泄露：强制不缓存任何搜索请求结果
    proxy_buffering off;
    proxy_cache off;

    location / {
        proxy_pass http://127.0.0.1:8080;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;

        # 提高连接与读取超时，防止上游多源聚合慢响应导致 504 Gateway Timeout
        proxy_connect_timeout 15s;
        proxy_read_timeout 60s;
        proxy_send_timeout 60s;
    }
}

sudo nginx -t && sudo systemctl reload nginx

# 查询 "Linux BBR 加速" 并输出前 3 条结构化 JSON
curl -s "https://search.0000996.xyz/search?q=Linux+BBR+加速&format=json" | jq '.results[0:3][] | {title: .title, url: .url, content: .content}'

{
  "title": "VPS 开启 BBR 网络加速指南",
  "url": "https://blog.0000996.xyz/archives/vps-bbr-guide/",
  "content": "详解 Linux 内核原生 BBR 拥塞控制算法部署，彻底解决海外节点网络丢包与吞吐低下..."
}

import requests

def private_web_search(query: str, max_results: int = 5):
    url = "https://search.0000996.xyz/search"
    params = {
        "q": query,
        "format": "json",
        "language": "zh-CN"
    }
    resp = requests.get(url, params=params, timeout=10)
    data = resp.json()

    # 提取聚合后的结果
    return [
        {"title": item.get("title"), "link": item.get("url"), "snippet": item.get("content")}
        for item in data.get("results", [])[:max_results]
    ]

# 实测调用
results = private_web_search("最新开源 AI 智能体框架")
for r in results:
    print(f"[{r['title']}] -> {r['link']}")