/**
 * Social Profile Finder
 * 
 * Finds company + decision-maker social profiles:
 * - Instagram (business account)
 * - Facebook (business page)
 * - Twitter/X
 * - YouTube
 * 
 * Two sources:
 * 1. Website footer/header scraping (most reliable)
 * 2. Google search fallback
 * 
 * Phase 2 uses these for multi-channel outreach.
 */

import { chromium } from "playwright";
import { playwrightLimiter } from "../../shared/utils/rate-limiter";
import { serperLimiter } from "../../shared/utils/rate-limiter";
import { logger } from "../../shared/utils/logger";
import axios from "axios";
import { getEnv } from "../../shared/config/env";

export interface SocialProfiles {
  instagram: string | null;
  facebook: string | null;
  twitter: string | null;
  youtube: string | null;
  source: "website" | "google" | "mixed";
}

/**
 * Find all social profiles for a company.
 * Method 1 first (website scrape), then Google fills gaps.
 */
export async function findSocialProfiles(
  domain: string,
  companyName: string,
  websiteHtml?: string
): Promise<SocialProfiles> {
  const profiles: SocialProfiles = {
    instagram: null,
    facebook: null,
    twitter: null,
    youtube: null,
    source: "website",
  };

  // ── Method 1: Extract from website HTML ────────────────────
  if (websiteHtml) {
    extractFromHtml(websiteHtml, profiles);
  } else {
    // Scrape website specifically for social links
    await scrapeWebsiteForSocials(domain, profiles);
  }

  // ── Method 2: Google search for missing profiles ───────────
  const missing = getMissing(profiles);
  if (missing.length > 0) {
    await searchGoogleForSocials(companyName, domain, profiles, missing);
    if (profiles.source === "website" && missing.some(p => profiles[p as keyof SocialProfiles])) {
      profiles.source = "mixed";
    }
  }

  const found = [profiles.instagram, profiles.facebook, profiles.twitter, profiles.youtube]
    .filter(Boolean).length;
  logger.info({ domain, found }, "Social profiles discovered");

  return profiles;
}

// ─── Method 1: HTML extraction ──────────────────────────────

const SOCIAL_PATTERNS = {
  instagram: /https?:\/\/(www\.)?instagram\.com\/[a-zA-Z0-9._]+/gi,
  facebook: /https?:\/\/(www\.)?(facebook|fb)\.com\/[a-zA-Z0-9.]+/gi,
  twitter: /https?:\/\/(www\.)?(twitter|x)\.com\/[a-zA-Z0-9_]+/gi,
  youtube: /https?:\/\/(www\.)?youtube\.com\/(channel|c|@)[\/a-zA-Z0-9._-]+/gi,
};

function extractFromHtml(html: string, profiles: SocialProfiles): void {
  for (const [platform, pattern] of Object.entries(SOCIAL_PATTERNS)) {
    const matches = html.match(pattern);
    if (matches && matches.length > 0) {
      // Take first match, clean it
      const url = cleanSocialUrl(matches[0], platform);
      if (url && !isGenericSocial(url)) {
        (profiles as Record<string, unknown>)[platform] = url;
      }
    }
  }
}

// ─── Website scrape (if HTML not already available) ──────────

async function scrapeWebsiteForSocials(domain: string, profiles: SocialProfiles): Promise<void> {
  try {
    await playwrightLimiter.consume("playwright");

    const browser = await chromium.launch({ headless: true, args: ["--no-sandbox"] });
    const context = await browser.newContext({
      userAgent: "Mozilla/5.0 (compatible; ResearchBot/1.0)",
    });
    const page = await context.newPage();

    await page.goto(`https://${domain}`, { waitUntil: "domcontentloaded", timeout: 12_000 });
    
    // Get all link hrefs on the page
    const links = await page.$$eval("a[href]", (anchors) =>
      anchors.map((a) => a.getAttribute("href") ?? "")
    );

    const pageHtml = links.join("\n");
    extractFromHtml(pageHtml, profiles);

    await page.close();
    await context.close();
    await browser.close();
  } catch (err) {
    logger.debug({ domain, err }, "Social scrape failed — trying Google");
  }
}

// ─── Method 2: Google search ────────────────────────────────

async function searchGoogleForSocials(
  companyName: string,
  domain: string,
  profiles: SocialProfiles,
  missing: string[]
): Promise<void> {
  const searchMap: Record<string, string> = {
    instagram: `"${companyName}" site:instagram.com`,
    facebook: `"${companyName}" site:facebook.com`,
    twitter: `"${companyName}" site:twitter.com OR site:x.com`,
    youtube: `"${companyName}" site:youtube.com`,
  };

  for (const platform of missing) {
    try {
      await serperLimiter.consume("serper");

      const env = getEnv();
      const response = await axios.post(
        "https://google.serper.dev/search",
        { q: searchMap[platform], num: 3 },
        {
          headers: {
            "X-API-KEY": env.SERPER_API_KEY,
            "Content-Type": "application/json",
          },
          timeout: 6_000,
        }
      );

      const organic = response.data?.organic ?? [];
      for (const result of organic) {
        const url = cleanSocialUrl(result.link, platform);
        if (url && !isGenericSocial(url)) {
          // Verify it mentions company name or domain in snippet
          const snippet = (result.snippet ?? "").toLowerCase();
          const title = (result.title ?? "").toLowerCase();
          const combined = `${snippet} ${title}`;
          
          const companyWords = companyName.toLowerCase().split(/\s+/);
          const hasCompany = companyWords.some(w => w.length > 3 && combined.includes(w));
          
          if (hasCompany || combined.includes(domain.replace(/\.\w+$/, ""))) {
            (profiles as Record<string, unknown>)[platform] = url;
            break;
          }
        }
      }
    } catch (err) {
      logger.debug({ platform, err }, "Social Google search failed — skipping");
    }
  }
}

// ─── Helpers ─────────────────────────────────────────────────

function getMissing(profiles: SocialProfiles): string[] {
  return ["instagram", "facebook", "twitter", "youtube"]
    .filter(p => !(profiles as Record<string, unknown>)[p]);
}

function cleanSocialUrl(url: string, platform: string): string | null {
  try {
    const parsed = new URL(url);
    // Remove query params and fragments
    return `${parsed.protocol}//${parsed.hostname}${parsed.pathname.replace(/\/$/, "")}`;
  } catch {
    return null;
  }
}

function isGenericSocial(url: string): boolean {
  // Filter out generic profile links (not actual company pages)
  const genericPaths = ["/share", "/sharer", "/login", "/signup", "/help", "/about", "/policies"];
  return genericPaths.some(p => url.includes(p));
}