scripts/build_static_site.py

"""Reproducibly build the atendyo.com STATIC site (Netlify) from source.

atendyo.com is served statically by Netlify (the backend only renders these same
pages). This regenerates the EXACT file tree the site needs so a landing/blog
change can be redeployed without hand-assembling files in /tmp:

  index.html              <- app/landing_ui/index.html (verbatim)
  og.png                  <- app/landing_ui/og.png
  blog/index.html         <- render_index("es")
  blog/en.html            <- render_index("en")
  blog/<slug>.html        <- render_article(post) for every post
  robots.txt, sitemap.xml <- regenerated, canonical on atendyo.com
  _redirects              <- the /demo -> backend redirect (load-bearing: every
                             cold-email demo link is atendyo.com/demo?d=...)

For static hosting the EN blog index lives at /blog/en (file blog/en.html), so we
rewrite the renderer's "/blog?lang=en" links to "/blog/en".

Usage: python -m scripts.build_static_site [output_dir]   (default: dist/site)
"""
from __future__ import annotations

import shutil
import sys
from pathlib import Path

ROOT = Path(__file__).resolve().parent.parent
sys.path.insert(0, str(ROOT))

from app.blog.posts import POSTS  # noqa: E402
from app.blog.render import render_article, render_index  # noqa: E402

SITE = "https://atendyo.com"
BACKEND = "https://victor34593993-flexigo-support-bot.hf.space"
LANDING = ROOT / "app" / "landing_ui" / "index.html"
OG = ROOT / "app" / "landing_ui" / "og.png"
LEGAL_DIR = ROOT / "app" / "legal_ui"
# Served at /legal/<page> (pretty URL). The privacy policy URL Meta needs is
# https://atendyo.com/legal/privacidad — must be a real, accessible page.
LEGAL_PAGES = ("aviso-legal", "privacidad", "cookies", "condiciones", "encargo")


def _static(html: str) -> str:
    # the EN blog index is a file (/blog/en), not a query string, when static
    return html.replace(f"{SITE}/blog?lang=en", f"{SITE}/blog/en")


def _robots() -> str:
    return (
        "User-agent: *\n"
        "Allow: /$\n"
        "Allow: /blog\n"
        "Disallow: /admin\n"
        "Disallow: /portal\n"
        "Disallow: /preview\n"
        "Disallow: /chat\n"
        "Disallow: /widget-config\n"
        f"Sitemap: {SITE}/sitemap.xml\n"
    )


def _sitemap() -> str:
    rows = [
        f"  <url><loc>{SITE}/</loc><changefreq>weekly</changefreq><priority>1.0</priority></url>",
        f"  <url><loc>{SITE}/blog</loc><changefreq>weekly</changefreq><priority>0.7</priority></url>",
    ]
    for p in POSTS:
        rows.append(
            f"  <url><loc>{SITE}/blog/{p.slug}</loc><lastmod>{p.date}</lastmod>"
            "<changefreq>monthly</changefreq><priority>0.6</priority></url>"
        )
    return (
        '<?xml version="1.0" encoding="UTF-8"?>\n'
        '<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">\n'
        + "\n".join(rows) + "\n</urlset>\n"
    )


def _redirects() -> str:
    # The ONLY redirect the site needs: atendyo.com/demo -> backend /demo (302,
    # query string auto-forwarded). Load-bearing for every email demo link.
    # MUST be a _redirects file in the publish root — a netlify.toml inside the
    # deployed --dir is published as a plain file and NOT processed as config.
    return f"/demo  {BACKEND}/demo  302\n"


def build(out: Path) -> None:
    if out.exists():
        shutil.rmtree(out)
    blog = out / "blog"
    blog.mkdir(parents=True)

    (out / "index.html").write_text(LANDING.read_text(encoding="utf-8"), encoding="utf-8")
    shutil.copyfile(OG, out / "og.png")

    legal = out / "legal"
    legal.mkdir()
    for page in LEGAL_PAGES:
        (legal / f"{page}.html").write_text(
            (LEGAL_DIR / f"{page}.html").read_text(encoding="utf-8"), encoding="utf-8")

    (blog / "index.html").write_text(_static(render_index("es")), encoding="utf-8")
    (blog / "en.html").write_text(_static(render_index("en")), encoding="utf-8")
    for p in POSTS:
        (blog / f"{p.slug}.html").write_text(_static(render_article(p)), encoding="utf-8")

    (out / "robots.txt").write_text(_robots(), encoding="utf-8")
    (out / "sitemap.xml").write_text(_sitemap(), encoding="utf-8")
    (out / "_redirects").write_text(_redirects(), encoding="utf-8")

    n = len(list(blog.glob("*.html")))
    print(f"built site at {out}: index + {n} blog pages + {len(LEGAL_PAGES)} legal "
          "pages + sitemap/robots/_redirects/og.png")


if __name__ == "__main__":
    target = Path(sys.argv[1]) if len(sys.argv) > 1 else ROOT / "dist" / "site"
    build(target)