#!/usr/bin/env python3
import requests
import datetime
import html
import xml.etree.ElementTree as ET
from pathlib import Path

FEED_URL = "https://www.theatlantic.com/feed/author/charlie-warzel/"
CATALOG_PATH = Path("catalog.html")
MAX_ITEMS = 50  # tweak if you want more or fewer


def fetch_feed(url: str) -> str:
    resp = requests.get(url, timeout=15)
    resp.raise_for_status()
    return resp.text


def parse_items(feed_xml: str):
    """
    Return a list of dicts: {title, link, date}
    Date is YYYY-MM-DD (or "" if missing).
    """
    root = ET.fromstring(feed_xml)
    # RSS feeds usually: <rss><channel><item>…</item></channel></rss>
    channel = root.find("channel")
    if channel is None:
        return []

    items = []
    for item in channel.findall("item"):
        title_el = item.find("title")
        link_el = item.find("link")
        pub_el = item.find("pubDate")

        title = title_el.text.strip() if title_el is not None and title_el.text else ""
        link = link_el.text.strip() if link_el is not None and link_el.text else ""
        pub_raw = pub_el.text.strip() if pub_el is not None and pub_el.text else ""

        # Try to parse pubDate -> YYYY-MM-DD
        date_str = ""
        if pub_raw:
            try:
                # Example format: "Wed, 20 Nov 2024 10:00:00 -0400"
                dt = datetime.datetime.strptime(pub_raw, "%a, %d %b %Y %H:%M:%S %z")
                date_str = dt.strftime("%Y-%m-%d")
            except Exception:
                # If parsing fails, just leave date_str empty
                date_str = ""

        if title and link:
            items.append(
                {
                    "title": title,
                    "link": link,
                    "date": date_str,
                }
            )

    return items


def build_list_html(items):
    """
    Build the HTML for the list of posts.
    We'll produce something like:

      <h2>From The Atlantic</h2>
      <ul class="post-list">
        <li>
          <a href="...">Title</a>
          <span class="date">— <span class="meta">YYYY-MM-DD</span></span>
        </li>
        ...
      </ul>
    """
    parts = []
    parts.append('      <h2>From The Atlantic</h2>')
    parts.append('      <ul class="post-list">')

    for item in items[:MAX_ITEMS]:
        title = html.escape(item["title"])
        link = html.escape(item["link"])
        date = item["date"] or ""
        if date:
            date_html = f'<span class="date">— <span class="meta">{date}</span></span>'
        else:
            date_html = ""

        li = f'        <li>\n          <a href="{link}">{title}</a> {date_html}\n        </li>'
        parts.append(li)

    parts.append("      </ul>")
    return "\n".join(parts)


def update_catalog():
    if not CATALOG_PATH.exists():
        raise SystemExit(f"catalog.html not found at {CATALOG_PATH.resolve()}")

    print("Fetching feed…")
    xml_text = fetch_feed(FEED_URL)
    items = parse_items(xml_text)
    if not items:
        raise SystemExit("No items found in feed; not touching catalog.html.")

    list_html = build_list_html(items)

    print("Reading catalog.html…")
    original = CATALOG_PATH.read_text(encoding="utf-8")

    # We’ll look for the block starting with <h2> and <ul class="post-list">,
    # and replace it entirely.
    marker_start = "<h2>From The Atlantic</h2>"
    marker_ul = '<ul class="post-list">'

    if marker_start in original:
        # Replace from marker_start to end of the </ul> that follows
        before, _, rest = original.partition(marker_start)
        _, ul_start, rest2 = rest.partition(marker_ul)

        # rest2 now starts just after the opening <ul class="post-list">
        # we need to find the closing </ul> that matches.
        ul_close = "</ul>"
        ul_close_idx = rest2.find(ul_close)
        if ul_close_idx == -1:
            raise SystemExit("Could not find closing </ul> after From The Atlantic block.")

        after = rest2[ul_close_idx + len(ul_close):]
        new_content = before + list_html + after
    else:
        # If marker not found, try to find the first <h1>Catalog</h1> and insert after it.
        marker_catalog = "<h1>Catalog</h1>"
        if marker_catalog not in original:
            raise SystemExit("Could not find <h1>Catalog</h1> in catalog.html")

        before, _, after = original.partition(marker_catalog)
        # keep the catalog heading and meta text as-is, and inject list_html after that block's metadata
        # e.g., we expect something like:
        #   <h1>Catalog</h1>
        #   <p class="meta">…</p>
        # so we place our list_html after that.
        meta_marker = "</p>"
        if meta_marker in after:
            meta_before, meta_close, meta_after = after.partition(meta_marker)
            # include the closing </p> then insert our list_html
            new_after = meta_before + meta_close + "\n\n" + list_html + meta_after
            new_content = before + marker_catalog + new_after
        else:
            # Fallback: just stick the list_html right after <h1>Catalog</h1>
            new_content = before + marker_catalog + "\n" + list_html + after

    print("Writing updated catalog.html…")
    CATALOG_PATH.write_text(new_content, encoding="utf-8")
    print("Done. Now commit and push the changes to GitHub.")


if __name__ == "__main__":
  update_catalog()