o2o-infinith-demo/scripts/import-registry.ts

/**
 * Import clinic_registry_working.csv into clinic_registry table.
 *
 * Usage:
 *   npx tsx scripts/import-registry.ts
 *
 * Requires env vars:
 *   SUPABASE_URL, SUPABASE_SERVICE_ROLE_KEY
 *
 * Or use .env.local file.
 */

import { createClient } from "@supabase/supabase-js";
import { readFileSync } from "fs";
import { config } from "dotenv";

config({ path: ".env.local" });

const SUPABASE_URL = process.env.SUPABASE_URL || process.env.VITE_SUPABASE_URL;
const SUPABASE_KEY = process.env.SUPABASE_SERVICE_ROLE_KEY;

if (!SUPABASE_URL || !SUPABASE_KEY) {
  console.error("Missing SUPABASE_URL or SUPABASE_SERVICE_ROLE_KEY");
  process.exit(1);
}

const supabase = createClient(SUPABASE_URL, SUPABASE_KEY);

// CSV column indices (0-based)
const COL = {
  name: 0,
  brand_group: 1,
  district: 2,
  branches: 3,
  website_kr: 4,
  website_en: 5,
  youtube_url: 6,
  // 7: youtube_note
  instagram_kr_url: 8,
  // 9: instagram_kr_note
  instagram_en_url: 10,
  // 11: instagram_en_note
  facebook_url: 12,
  // 13: facebook_note
  tiktok_url: 14,
  // 15: tiktok_note
  gangnam_unni_url: 16,
  // 17: gangnam_unni_note
  naver_blog_url: 18,
  // 19: naver_blog_note
  naver_place_url: 20,
  // 21: naver_place_reviews_note
  google_maps_url: 22,
  // 23: google_reviews_note
} as const;

function extractDomain(url: string): string {
  try {
    return new URL(url).hostname.replace(/^www\./, "");
  } catch {
    // Handle URLs without protocol
    const clean = url.replace(/^https?:\/\//, "").replace(/^www\./, "");
    return clean.split("/")[0];
  }
}

function parseCSVLine(line: string): string[] {
  // Simple CSV parser (no quoted fields with commas in this CSV)
  return line.split(",");
}

async function main() {
  const csv = readFileSync("data/clinic-registry/clinic_registry_working.csv", "utf8");
  const lines = csv.split("\n").filter((l) => l.trim());
  const rows = lines.slice(1); // skip header

  console.log(`Parsing ${rows.length} clinics from CSV...`);

  const records: Record<string, unknown>[] = [];
  const skipped: string[] = [];

  for (const line of rows) {
    const cols = parseCSVLine(line);
    const name = cols[COL.name]?.trim();
    const website = cols[COL.website_kr]?.trim();

    if (!name || !website) {
      skipped.push(name || "(unnamed)");
      continue;
    }

    const domain = extractDomain(website);
    if (!domain) {
      skipped.push(name);
      continue;
    }

    records.push({
      name,
      name_aliases: [], // Can be enriched later
      domain,
      website_url: website,
      brand_group: cols[COL.brand_group]?.trim() || null,
      district: cols[COL.district]?.trim() || null,
      branches: cols[COL.branches]?.trim() || null,
      website_en: cols[COL.website_en]?.trim() || null,
      youtube_url: cols[COL.youtube_url]?.trim() || null,
      instagram_url: cols[COL.instagram_kr_url]?.trim() || null,
      instagram_en_url: cols[COL.instagram_en_url]?.trim() || null,
      facebook_url: cols[COL.facebook_url]?.trim() || null,
      tiktok_url: cols[COL.tiktok_url]?.trim() || null,
      gangnam_unni_url: cols[COL.gangnam_unni_url]?.trim() || null,
      naver_blog_url: cols[COL.naver_blog_url]?.trim() || null,
      naver_place_url: cols[COL.naver_place_url]?.trim() || null,
      google_maps_url: cols[COL.google_maps_url]?.trim() || null,
      verified_by: "scrape",
      is_active: true,
    });
  }

  console.log(`Prepared ${records.length} records (skipped ${skipped.length}: ${skipped.join(", ")})`);

  // Upsert in batches of 20
  const BATCH_SIZE = 20;
  let inserted = 0;
  let updated = 0;
  let errors = 0;

  for (let i = 0; i < records.length; i += BATCH_SIZE) {
    const batch = records.slice(i, i + BATCH_SIZE);
    const { data, error } = await supabase
      .from("clinic_registry")
      .upsert(batch, { onConflict: "domain" })
      .select("id, domain");

    if (error) {
      console.error(`Batch ${i / BATCH_SIZE + 1} error:`, error.message);
      errors += batch.length;
    } else {
      inserted += data?.length || 0;
      console.log(`Batch ${i / BATCH_SIZE + 1}: ${data?.length} rows upserted`);
    }
  }

  console.log(`\nDone! Inserted/updated: ${inserted}, Errors: ${errors}`);

  // Verify
  const { count } = await supabase
    .from("clinic_registry")
    .select("*", { count: "exact", head: true });
  console.log(`Total rows in clinic_registry: ${count}`);
}

main().catch(console.error);