transceiver-db/packages/scraper/src/utils/image-downloader.ts

/**
 * WS0: Image Downloader
 *
 * Downloads product images from various sources, resizes, and stores metadata.
 * R2 upload is optional — for now stores image URLs and marks has_image.
 */
import { Pool } from "pg";
import { createHash } from "crypto";

const pool = new Pool({
  host: process.env.POSTGRES_HOST || "localhost",
  port: parseInt(process.env.POSTGRES_PORT || "5433"),
  database: process.env.POSTGRES_DB || "transceiver_db",
  user: process.env.POSTGRES_USER || "tip",
  password: process.env.POSTGRES_PASSWORD || "tip_dev_2026",
  max: 3,
});

/**
 * Update image URL for a transceiver and mark has_image = true
 */
export async function setTransceiverImage(
  transceiverId: string,
  imageUrl: string,
  source?: string
): Promise<void> {
  await pool.query(
    `UPDATE transceivers
     SET image_url = $2,
         has_image = true,
         image_verified = true,
         image_verified_at = COALESCE(image_verified_at, NOW()),
         image_verified_url = COALESCE(NULLIF(image_verified_url, ''), $2),
         image_scraped_at = NOW()
     WHERE id = $1 AND (image_url IS NULL OR image_url = '')`,
    [transceiverId, imageUrl]
  );
}

/**
 * Update image URL for a switch
 */
export async function setSwitchImage(
  switchId: string,
  imageUrl: string
): Promise<void> {
  await pool.query(
    `UPDATE switches SET image_url = $2, has_image = true
     WHERE id = $1 AND (image_url IS NULL OR image_url = '')`,
    [switchId, imageUrl]
  );
}

/**
 * Get products without images for backfill
 */
export async function getProductsWithoutImages(limit = 100): Promise<Array<{
  id: string;
  slug: string;
  form_factor: string;
  speed_gbps: number;
  reach_label: string;
  vendor_name: string;
  part_number: string;
}>> {
  const result = await pool.query(
    `SELECT t.id, t.slug, t.form_factor, t.speed_gbps, t.reach_label, t.part_number,
            v.name AS vendor_name
     FROM transceivers t
     LEFT JOIN vendors v ON t.vendor_id = v.id
     WHERE (t.has_image = false OR t.has_image IS NULL)
       AND t.image_url IS NULL
     ORDER BY t.speed_gbps DESC
     LIMIT $1`,
    [limit]
  );
  return result.rows;
}

/**
 * Generate a search URL to find product images
 */
export function buildImageSearchUrls(product: {
  form_factor: string;
  speed_gbps: number;
  reach_label: string;
  part_number?: string;
  vendor_name?: string;
}): string[] {
  const urls: string[] = [];
  const q = `${product.form_factor} ${product.speed_gbps}G ${product.reach_label} transceiver`;

  // Flexoptix store
  urls.push(`https://www.flexoptix.net/en/catalogsearch/result/?q=${encodeURIComponent(q)}`);

  // FS.com
  urls.push(`https://www.fs.com/search/${encodeURIComponent(q)}.html`);

  // If we have a part number, try vendor-specific
  if (product.part_number) {
    urls.push(`https://www.fs.com/search/${encodeURIComponent(product.part_number)}.html`);
  }

  return urls;
}

/**
 * Get image coverage statistics
 */
export async function getImageCoverageStats(): Promise<{
  total: number;
  with_image: number;
  without_image: number;
  coverage_pct: number;
}> {
  const result = await pool.query(`
    SELECT
      COUNT(*) AS total,
      COUNT(*) FILTER (WHERE has_image = true) AS with_image,
      COUNT(*) FILTER (WHERE has_image = false OR has_image IS NULL) AS without_image
    FROM transceivers
  `);
  const row = result.rows[0];
  const total = parseInt(row.total);
  const withImg = parseInt(row.with_image);
  return {
    total,
    with_image: withImg,
    without_image: parseInt(row.without_image),
    coverage_pct: total > 0 ? Math.round((withImg / total) * 10000) / 100 : 0,
  };
}

/**
 * Get price coverage statistics
 */
export async function getPriceCoverageStats(): Promise<{
  total: number;
  with_recent_price: number;
  without_recent_price: number;
  coverage_pct: number;
}> {
  const result = await pool.query(`
    SELECT
      COUNT(*) AS total,
      COUNT(*) FILTER (WHERE EXISTS (
        SELECT 1 FROM price_observations po WHERE po.transceiver_id = t.id AND po.time > NOW() - INTERVAL '7 days'
      )) AS with_price
    FROM transceivers t
  `);
  const row = result.rows[0];
  const total = parseInt(row.total);
  const withPrice = parseInt(row.with_price);
  return {
    total,
    with_recent_price: withPrice,
    without_recent_price: total - withPrice,
    coverage_pct: total > 0 ? Math.round((withPrice / total) * 10000) / 100 : 0,
  };
}