kmtnflower/backend/app/sitemap_export.py

"""Sitemap & robots.txt export: generate tenant-aware SEO files.

Because this is a white-label app where each tenant has its own domain
(set via KMTN_WEBSITE_URL), the sitemap and robots.txt are generated
at runtime from the station config — one static write per startup,
plus an optional admin-triggered regeneration.

Follows the same pattern as theme_export.py:
- Atomic write (tmp + os.replace)
- Written to STATIC_FILES_DIR so Nginx serves them
- Uses WEBSITE_URL for per-tenant domain resolution
"""

import logging
from datetime import datetime, timezone
from pathlib import Path

from app.config import settings

logger = logging.getLogger(__name__)

# Public routes included in the sitemap with their change frequency and priority.
SEOMAP_ROUTES = [
    {"path": "/", "changefreq": "daily", "priority": 1.0},
    {"path": "/about", "changefreq": "weekly", "priority": 0.8},
    {"path": "/schedule", "changefreq": "weekly", "priority": 0.8},
    {"path": "/events", "changefreq": "daily", "priority": 0.8},
    {"path": "/contact", "changefreq": "monthly", "priority": 0.5},
]


def build_sitemap_xml(base_url: str) -> str:
    """Build a valid sitemap.xml string for the given base URL."""
    url = base_url.rstrip("/")
    entries = []
    for route in SEOMAP_ROUTES:
        entries.append(
            f"  <url>\n"
            f"    <loc>{url}{route['path']}</loc>\n"
            f"    <changefreq>{route['changefreq']}</changefreq>\n"
            f"    <priority>{route['priority']}</priority>\n"
            f"  </url>"
        )
    return (
        '<?xml version="1.0" encoding="UTF-8"?>\n'
        '<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">\n'
        + "\n".join(entries) + "\n"
        "</urlset>\n"
    )


def build_robots_txt(sitemap_url: str) -> str:
    """Build a robots.txt string that references the correct sitemap URL."""
    return (
        "User-agent: *\n"
        "Allow: /\n"
        "\n"
        "Disallow: /admin\n"
        "Disallow: /login\n"
        "\n"
        f"Sitemap: {sitemap_url}/sitemap.xml\n"
    )


def get_sitemap_status() -> dict:
    """Check if sitemap.xml and robots.txt exist on disk and return metadata."""
    base = Path(settings.STATIC_FILES_DIR)
    result = {}
    for name in ("sitemap.xml", "robots.txt"):
        path = base / name
        if not path.exists():
            result[name] = {"exists": False, "path": str(path)}
            continue
        stat = path.stat()
        result[name] = {
            "exists": True,
            "path": str(path),
            "size": stat.st_size,
            "modified": datetime.fromtimestamp(stat.st_mtime, tz=timezone.utc).isoformat(),
        }
    return result


def write_sitemap_files() -> dict:
    """Write sitemap.xml and robots.txt to the static files directory.

    Uses atomic write (write to .tmp, then os.replace) to prevent
    Nginx from serving a partially-written file.

    Returns a status dict with file paths.
    """
    base_url = settings.WEBSITE_URL
    static_dir = Path(settings.STATIC_FILES_DIR)
    static_dir.mkdir(parents=True, exist_ok=True)

    # sitemap.xml
    sitemap_content = build_sitemap_xml(base_url)
    sitemap_path = static_dir / "sitemap.xml"
    tmp_path = sitemap_path.with_suffix(".tmp")
    tmp_path.write_text(sitemap_content, encoding="utf-8")
    tmp_path.replace(sitemap_path)
    logger.info("Wrote sitemap.xml (%d bytes) to %s, base_url=%s", len(sitemap_content), sitemap_path, base_url)

    # robots.txt
    robots_content = build_robots_txt(base_url)
    robots_path = static_dir / "robots.txt"
    tmp_path = robots_path.with_suffix(".tmp")
    tmp_path.write_text(robots_content, encoding="utf-8")
    tmp_path.replace(robots_path)
    logger.info("Wrote robots.txt (%d bytes) to %s", len(robots_content), robots_path)

    return {
        "sitemap.xml": str(sitemap_path),
        "robots.txt": str(robots_path),
        "base_url": base_url,
    }