import { Logger } from "../utils/Logger";
import { createTurndownService } from "./htmlExtractor";

export interface ExtractedRssContent {
  title?: string;
  author?: string;
  publishedDate?: string;
  markdown: string;
}

/**
 * Detects RSS or Atom feed link advertised in the document <head>.
 */
export function detectRssFeedUrl(doc: Document, baseUrl: string): string | undefined {
  const selectors = [
    'link[rel="alternate"][type="application/rss+xml"]',
    'link[rel="alternate"][type="application/atom+xml"]',
  ];

  for (const selector of selectors) {
    const link = doc.querySelector(selector);
    if (link) {
      const href = link.getAttribute("href");
      if (href && href.trim()) {
        try {
          return new URL(href.trim(), baseUrl).href;
        } catch {
          // Ignore invalid URL resolution
        }
      }
    }
  }

  return undefined;
}

/**
 * Lightweight RSS/Atom XML string parser.
 * Extracts matching item/entry based on targetUrl and converts its HTML content to Markdown.
 */
export function parseRssFeed(xmlText: string, targetUrl: string): ExtractedRssContent | null {
  if (!xmlText || !xmlText.trim()) return null;

  try {
    // Helper regex extractors
    const getTagContent = (xmlFragment: string, tagName: string): string | undefined => {
      const escapedTag = tagName.replace(":", "\\:");
      const regex = new RegExp(`<${escapedTag}(?:\\s+[^>]*)?>([\\s\\S]*?)</${escapedTag}>`, "i");
      const match = xmlFragment.match(regex);
      if (!match) return undefined;
      let val = match[1].trim();
      // Unwrap CDATA if present
      const cdataMatch = val.match(/^<!\[CDATA\[([\s\S]*?)\]\]>$/i);
      if (cdataMatch) {
        val = cdataMatch[1].trim();
      }
      return val || undefined;
    };

    const getAtomLink = (entryFragment: string): string | undefined => {
      const match = entryFragment.match(/<link(?:\s+[^>]*)?href=["']([^"']+)["']/i);
      return match ? match[1].trim() : undefined;
    };

    // Extract item (RSS 2.0) or entry (Atom) blocks
    const isAtom = xmlText.includes("<entry") || xmlText.includes("http://www.w3.org/2005/Atom");
    const itemRegex = isAtom ? /<entry(?:\s+[^>]*)?>([\s\S]*?)<\/entry>/gi : /<item(?:\s+[^>]*)?>([\s\S]*?)<\/item>/gi;

    const items: { fragment: string; link?: string }[] = [];
    let match: RegExpExecArray | null;

    while ((match = itemRegex.exec(xmlText)) !== null) {
      const fragment = match[1];
      const link = isAtom
        ? getAtomLink(fragment) || getTagContent(fragment, "id")
        : getTagContent(fragment, "link") || getTagContent(fragment, "guid");
      items.push({ fragment, link });
    }

    if (items.length === 0) return null;

    // Target URL matching
    let targetItem = items.find((item) => {
      if (!item.link) return false;
      try {
        const itemUrl = new URL(item.link, targetUrl).href;
        const targetObj = new URL(targetUrl).href;
        return itemUrl === targetObj || item.link.includes(targetUrl) || targetUrl.includes(item.link);
      } catch {
        return item.link === targetUrl;
      }
    });

    // Fallback to first item if no direct URL match
    if (!targetItem) {
      targetItem = items[0];
    }

    const frag = targetItem.fragment;
    const title = getTagContent(frag, "title");
    const author =
      getTagContent(frag, "dc:creator") ||
      getTagContent(frag, "author") ||
      getTagContent(frag, "name");
    const publishedDate =
      getTagContent(frag, "pubDate") ||
      getTagContent(frag, "published") ||
      getTagContent(frag, "updated") ||
      getTagContent(frag, "dc:date");
    const rawContentHtml =
      getTagContent(frag, "content:encoded") ||
      getTagContent(frag, "content") ||
      getTagContent(frag, "description") ||
      getTagContent(frag, "summary");

    if (!rawContentHtml) return null;

    const turndownService = createTurndownService();
    const markdown = turndownService.turndown(rawContentHtml).trim();

    if (!markdown) return null;

    return {
      title,
      author,
      publishedDate,
      markdown,
    };
  } catch (err) {
    Logger.debug(
      "[WebpageReader]",
      `Failed to parse RSS XML feed: ${err instanceof Error ? err.message : String(err)}`
    );
    return null;
  }
}

import { Logger } from "../utils/Logger";
import { createTurndownService } from "./htmlExtractor";

export interface ExtractedRssContent {
  title?: string;
  author?: string;
  publishedDate?: string;
  markdown: string;
}

/**
 * Detects RSS or Atom feed link advertised in the document <head>.
 */
export function detectRssFeedUrl(doc: Document, baseUrl: string): string | undefined {
  const selectors = [
    'link[rel="alternate"][type="application/rss+xml"]',
    'link[rel="alternate"][type="application/atom+xml"]',
  ];

  for (const selector of selectors) {
    const link = doc.querySelector(selector);
    if (link) {
      const href = link.getAttribute("href");
      if (href && href.trim()) {
        try {
          return new URL(href.trim(), baseUrl).href;
        } catch {
          // Ignore invalid URL resolution
        }
      }
    }
  }

  return undefined;
}

/**
 * Lightweight RSS/Atom XML string parser.
 * Extracts matching item/entry based on targetUrl and converts its HTML content to Markdown.
 */
export function parseRssFeed(xmlText: string, targetUrl: string): ExtractedRssContent | null {
  if (!xmlText || !xmlText.trim()) return null;

  try {
    // Helper regex extractors
    const getTagContent = (xmlFragment: string, tagName: string): string | undefined => {
      const escapedTag = tagName.replace(":", "\\:");
      const regex = new RegExp(`<${escapedTag}(?:\\s+[^>]*)?>([\\s\\S]*?)</${escapedTag}>`, "i");
      const match = xmlFragment.match(regex);
      if (!match) return undefined;
      let val = match[1].trim();
      // Unwrap CDATA if present
      const cdataMatch = val.match(/^<!\[CDATA\[([\s\S]*?)\]\]>$/i);
      if (cdataMatch) {
        val = cdataMatch[1].trim();
      }
      return val || undefined;
    };

    const getAtomLink = (entryFragment: string): string | undefined => {
      const match = entryFragment.match(/<link(?:\s+[^>]*)?href=["']([^"']+)["']/i);
      return match ? match[1].trim() : undefined;
    };

    // Extract item (RSS 2.0) or entry (Atom) blocks
    const isAtom = xmlText.includes("<entry") || xmlText.includes("http://www.w3.org/2005/Atom");
    const itemRegex = isAtom ? /<entry(?:\s+[^>]*)?>([\s\S]*?)<\/entry>/gi : /<item(?:\s+[^>]*)?>([\s\S]*?)<\/item>/gi;

    const items: { fragment: string; link?: string }[] = [];
    let match: RegExpExecArray | null;

    while ((match = itemRegex.exec(xmlText)) !== null) {
      const fragment = match[1];
      const link = isAtom
        ? getAtomLink(fragment) || getTagContent(fragment, "id")
        : getTagContent(fragment, "link") || getTagContent(fragment, "guid");
      items.push({ fragment, link });
    }

    if (items.length === 0) return null;

    // Target URL matching
    let targetItem = items.find((item) => {
      if (!item.link) return false;
      try {
        const itemUrl = new URL(item.link, targetUrl).href;
        const targetObj = new URL(targetUrl).href;
        return itemUrl === targetObj || item.link.includes(targetUrl) || targetUrl.includes(item.link);
      } catch {
        return item.link === targetUrl;
      }
    });

    // Fallback to first item if no direct URL match
    if (!targetItem) {
      targetItem = items[0];
    }

    const frag = targetItem.fragment;
    const title = getTagContent(frag, "title");
    const author =
      getTagContent(frag, "dc:creator") ||
      getTagContent(frag, "author") ||
      getTagContent(frag, "name");
    const publishedDate =
      getTagContent(frag, "pubDate") ||
      getTagContent(frag, "published") ||
      getTagContent(frag, "updated") ||
      getTagContent(frag, "dc:date");
    const rawContentHtml =
      getTagContent(frag, "content:encoded") ||
      getTagContent(frag, "content") ||
      getTagContent(frag, "description") ||
      getTagContent(frag, "summary");

    if (!rawContentHtml) return null;

    const turndownService = createTurndownService();
    const markdown = turndownService.turndown(rawContentHtml).trim();

    if (!markdown) return null;

    return {
      title,
      author,
      publishedDate,
      markdown,
    };
  } catch (err) {
    Logger.debug(
      "[WebpageReader]",
      `Failed to parse RSS XML feed: ${err instanceof Error ? err.message : String(err)}`
    );
    return null;
  }
}