lead-scraper/lib/services/googlemaps.ts

// Google Maps Places API (New)
// Docs: https://developers.google.com/maps/documentation/places/web-service/text-search
// Auth: X-Goog-Api-Key header
// Pricing: $32/1,000 requests — $200 free credit/month ≈ 6,250 free searches
//
// Text Search returns up to 20 results per page, max 3 pages (60 results) per query.
// Use multiple queries (different cities) to exceed 60 results.

import axios from "axios";
import { extractDomainFromUrl } from "@/lib/utils/domains";

const BASE_URL = "https://places.googleapis.com/v1/places:searchText";

// Fields we need — only request what we use to minimize billing
const FIELD_MASK = [
  "places.id",
  "places.displayName",
  "places.websiteUri",
  "places.formattedAddress",
  "places.nationalPhoneNumber",
  "places.businessStatus",
  "places.editorialSummary",
  "places.primaryTypeDisplayName",
  "nextPageToken",
].join(",");

export interface PlaceResult {
  placeId: string;
  name: string;
  website: string | null;
  domain: string | null;
  address: string;
  phone: string | null;
  description: string | null;
  category: string | null;
}

interface PlacesApiResponse {
  places?: Array<{
    id: string;
    displayName?: { text: string };
    websiteUri?: string;
    formattedAddress?: string;
    nationalPhoneNumber?: string;
    businessStatus?: string;
    editorialSummary?: { text: string };
    primaryTypeDisplayName?: { text: string };
  }>;
  nextPageToken?: string;
}

export async function searchPlaces(
  textQuery: string,
  apiKey: string,
  maxResults = 60,
  languageCode = "de"
): Promise<PlaceResult[]> {
  const results: PlaceResult[] = [];
  let pageToken: string | undefined;

  do {
    const body: Record<string, unknown> = {
      textQuery,
      pageSize: 20,
      languageCode,
    };
    if (pageToken) body.pageToken = pageToken;

    const response = await axios.post<PlacesApiResponse>(BASE_URL, body, {
      headers: {
        "X-Goog-Api-Key": apiKey,
        "X-Goog-FieldMask": FIELD_MASK,
        "Content-Type": "application/json",
      },
      timeout: 15000,
    });

    const data = response.data;
    const places = data.places || [];

    for (const place of places) {
      if (place.businessStatus && place.businessStatus !== "OPERATIONAL") continue;

      const website = place.websiteUri || null;
      const domain = website ? extractDomainFromUrl(website) : null;

      results.push({
        placeId: place.id,
        name: place.displayName?.text || "",
        website,
        domain,
        address: place.formattedAddress || "",
        phone: place.nationalPhoneNumber || null,
        description: place.editorialSummary?.text || null,
        category: place.primaryTypeDisplayName?.text || null,
      });
    }

    pageToken = data.nextPageToken;

    // Respect Google's recommendation: wait briefly between paginated requests
    if (pageToken && results.length < maxResults) {
      await sleep(500);
    }
  } while (pageToken && results.length < maxResults);

  return results.slice(0, maxResults);
}

/**
 * Search across multiple queries (e.g., different cities) and deduplicate by domain.
 */
export async function searchPlacesMultiQuery(
  queries: string[],
  apiKey: string,
  maxResultsPerQuery = 60,
  languageCode = "de",
  onProgress?: (done: number, total: number) => void
): Promise<PlaceResult[]> {
  const seenDomains = new Set<string>();
  const seenPlaceIds = new Set<string>();
  const allResults: PlaceResult[] = [];

  for (let i = 0; i < queries.length; i++) {
    try {
      const results = await searchPlaces(queries[i], apiKey, maxResultsPerQuery, languageCode);

      for (const r of results) {
        if (seenPlaceIds.has(r.placeId)) continue;
        if (r.domain && seenDomains.has(r.domain)) continue;

        seenPlaceIds.add(r.placeId);
        if (r.domain) seenDomains.add(r.domain);
        allResults.push(r);
      }

      onProgress?.(i + 1, queries.length);

      // Avoid hammering the API between queries
      if (i < queries.length - 1) await sleep(300);
    } catch (err) {
      console.error(`Google Maps search error for query "${queries[i]}":`, err);
    }
  }

  return allResults;
}

function sleep(ms: number) {
  return new Promise(r => setTimeout(r, ms));
}