域名迁移后的搜索引擎“饱和式”收录（bing推送一键脚本）

import os
import sys
import httpx
import xml.etree.ElementTree as ET
import time

# ================= 配置区域 =================
# 1. 基础路径配置
BASE_DIR = "/www/wwwroot/bing_push"
# 2. 你的新站 Sitemap 路径
SITEMAP_URL = "https://blog.oool.cc/sitemap.xml" 
# 3. 历史记录与运行日志
HISTORY_FILE = os.path.join(BASE_DIR, "pushed_history.txt")
LOG_FILE = os.path.join(BASE_DIR, "execution.log")
# 4. 关键凭证 (请务必脱敏替换)
INDEXNOW_KEY = "your_private_key_here" 
REAL_DOMAIN = "blog.oool.cc"
# ===========================================

class BingProManager:
    def __init__(self):
        self.endpoint = "https://www.bing.com/indexnow"

    def log(self, msg):
        """格式化日志记录"""
        timestamp = time.strftime('%Y-%m-%d %H:%M:%S')
        full_msg = f"[{timestamp}] {msg}"
        print(full_msg)
        if not os.path.exists(BASE_DIR): os.makedirs(BASE_DIR)
        with open(LOG_FILE, "a", encoding="utf-8") as f:
            f.write(full_msg + "\n")

    def get_urls(self):
        """解析 Sitemap 获取全量 URL 列表"""
        try:
            r = httpx.get(SITEMAP_URL, timeout=30.0)
            r.raise_for_status()
            root = ET.fromstring(r.content)
            # 命名空间处理，Sitemap 规范通常带有这个 schema
            ns = {'s': 'http://www.sitemaps.org/schemas/sitemap/0.9'}
            return [n.text for n in root.findall(".//s:loc", ns)]
        except Exception as e:
            self.log(f"❌ Sitemap 抓取/解析失败: {e}")
            return []

    def run_push(self):
        """执行 IndexNow 推送逻辑"""
        all_urls = self.get_urls()
        if not all_urls: return

        # 加载历史记录
        pushed = set()
        if os.path.exists(HISTORY_FILE):
            with open(HISTORY_FILE, "r") as f:
                pushed = {line.strip() for line in f}

        # 筛选从未推送过的新链接 (或迁移后的全量链接)
        new_urls = [u for u in all_urls if u not in pushed]
        
        if not new_urls:
            self.log("☕ 当前已是最新状态，无需推送。")
            return

        self.log(f"🚀 准备推送 {len(new_urls)} 条链接至 Bing...")
        
        # 构造 IndexNow 标准 Payload
        payload = {
            "host": REAL_DOMAIN,
            "key": INDEXNOW_KEY,
            "keyLocation": f"https://{REAL_DOMAIN}/{INDEXNOW_KEY}.txt",
            "urlList": new_urls
        }

        try:
            # 采用异步库 httpx 的同步模式，确保兼容性
            resp = httpx.post(self.endpoint, json=payload, timeout=20.0)
            if resp.status_code in [200, 202]:
                self.log(f"✅ 成功：Bing 接收成功。状态码: {resp.status_code}")
                # 写入历史文件，确保不重复推送
                with open(HISTORY_FILE, "a") as f:
                    for url in new_urls:
                        f.write(url + "\n")
            else:
                self.log(f"❌ 失败：状态码 {resp.status_code}，响应: {resp.text}")
        except Exception as e:
            self.log(f"❌ 网络请求异常: {e}")

    def show_history(self):
        """打印已成功推送的清单"""
        if os.path.exists(HISTORY_FILE):
            with open(HISTORY_FILE, "r") as f:
                print(f.read())
        else:
            print("⚠️ 暂无推送记录。")

if __name__ == "__main__":
    manager = BingProManager()
    # 简单的命令行参数控制
    if len(sys.argv) > 1:
        cmd = sys.argv[1].lower()
        if cmd == "history": manager.show_history()
        elif cmd == "logs":
            if os.path.exists(LOG_FILE):
                with open(LOG_FILE, "r") as f: print(f.read())
        else: manager.run_push()
    else:
        manager.run_push()

# 1. 确保环境依赖完整，防止服务器迁移或环境变动
python3 -m pip install httpx --break-system-packages > /dev/null 2>&1

# 2. 执行推送脚本
python3 /www/wwwroot/bing_push/bing_manager.py run