Sansheng-Intelligence/scripts/build_site.py

#!/usr/bin/env python3
"""构建可直接部署或由本地 HTTP 服务打开的独立静态站。"""

from __future__ import annotations

import argparse
import json
import os
import shutil
import subprocess
import sys
from pathlib import Path

from seo_config import BRAND_NAME, ENTITY_DEFINITION, KEYWORD_CLUSTERS, SECTION_INTENTS


ROOT = Path(__file__).resolve().parents[1]
PUBLISH_DATA = ("textile-intelligence.json", "textile-trends.json")
DEFAULT_SITE_URL = "http://127.0.0.1:8765"


def replace_directory(source: Path, target: Path) -> None:
    previous = target.with_name(f".{target.name}-previous")
    if previous.exists():
        shutil.rmtree(previous)
    if target.exists():
        target.replace(previous)
    try:
        source.replace(target)
    except Exception:
        if previous.exists() and not target.exists():
            previous.replace(target)
        raise
    if previous.exists():
        shutil.rmtree(previous)


def write_llms_file(target: Path, payload: dict, site_url: str) -> None:
    summary = payload.get("summary") or {}
    lines = [
        f"# {BRAND_NAME}",
        "",
        f"> {ENTITY_DEFINITION}",
        "",
        f"- 站点：{site_url}/",
        f"- 权威源：{payload.get('source_success', 0)}/{payload.get('source_target', 0)}",
        f"- 备货机会：{summary.get('calendar_featured', 0)}",
        f"- 展会：{summary.get('exhibitions', 0)}",
        f"- 政策：{summary.get('policies', 0)}",
        f"- 平台：{summary.get('platforms', 0)}",
        f"- 跨境活动：{summary.get('activities', 0)}",
        f"- 核心词：{'、'.join((*KEYWORD_CLUSTERS['core'], *KEYWORD_CLUSTERS['task']))}",
        "",
        "## 可索引栏目",
        "",
        f"- [备货日历]({site_url}/textile/calendar/)",
        f"- [展会通]({site_url}/textile/exhibitions/)",
        f"- [政策预警]({site_url}/textile/policies/)",
        f"- [趋势雷达]({site_url}/textile/standards/)",
        f"- [平台渠道]({site_url}/textile/platforms/)",
        f"- [跨境活动]({site_url}/textile/activities/)",
        "",
        "## 栏目回答意图",
        "",
        *[f"- {section}：{intent}" for section, intent in SECTION_INTENTS.items()],
        "",
        "官方事实、机器翻译与本站经营分析分区展示；机器翻译仅作阅读辅助，不替代报名、付款或合规提交。",
    ]
    (target / "llms.txt").write_text("\n".join(lines) + "\n", encoding="utf-8")


def build(data_dir: Path, output_dir: Path, site_url: str) -> None:
    data_dir = data_dir.resolve()
    output_dir = output_dir.resolve()
    staging = output_dir.with_name(f".{output_dir.name}-build")
    if staging.exists():
        shutil.rmtree(staging)
    shutil.copytree(ROOT / "public", staging)
    target_data = staging / "data"
    target_data.mkdir(parents=True)
    for name in PUBLISH_DATA:
        source = data_dir / name
        if not source.is_file():
            raise FileNotFoundError(f"缺少发布数据：{source}")
        shutil.copy2(source, target_data / name)

    index = staging / "index.html"
    html = index.read_text(encoding="utf-8")
    html = html.replace('<link rel="canonical" href="./" />', f'<link rel="canonical" href="{site_url}/" />')
    html = html.replace('<link rel="alternate" hreflang="zh-CN" href="./" />', f'<link rel="alternate" hreflang="zh-CN" href="{site_url}/" />')
    html = html.replace('<link rel="alternate" hreflang="x-default" href="./" />', f'<link rel="alternate" hreflang="x-default" href="{site_url}/" />')
    html = html.replace('<meta property="og:image" content="./assets/logo.svg" />', f'<meta property="og:image" content="{site_url}/assets/logo.svg" />')
    index.write_text(html, encoding="utf-8", newline="\n")

    subprocess.run([
        sys.executable,
        str(ROOT / "scripts" / "build_seo.py"),
        "--data", str(target_data / "textile-intelligence.json"),
        "--root", str(staging),
        "--site-url", site_url,
    ], check=True, cwd=ROOT)
    payload = json.loads((target_data / "textile-intelligence.json").read_text(encoding="utf-8"))
    write_llms_file(staging, payload, site_url)
    required = ("index.html", "robots.txt", "sitemap.xml", "llms.txt")
    missing = [name for name in required if not (staging / name).is_file()]
    if missing:
        raise RuntimeError(f"构建产物不完整：{missing}")
    replace_directory(staging, output_dir)


def main() -> int:
    parser = argparse.ArgumentParser(description="构建三笙AI情报局独立静态站")
    parser.add_argument("--data-dir", default=str(ROOT / "data"))
    parser.add_argument("--output-dir", default=str(ROOT / "dist"))
    parser.add_argument("--site-url", default=os.environ.get("SANSHENG_SITE_URL", DEFAULT_SITE_URL))
    args = parser.parse_args()
    site_url = args.site_url.rstrip("/")
    try:
        build(Path(args.data_dir), Path(args.output_dir), site_url)
    except Exception as exc:
        print(f"[构建失败] {exc}", file=sys.stderr)
        return 1
    print(f"[构建完成] {Path(args.output_dir).resolve()}")
    return 0


if __name__ == "__main__":
    raise SystemExit(main())