feat: expose hotel cost unit and iso country code via the api

src/BusProNet/Model/Hotel.php

@@ -35,4 +35,15 @@ class Hotel
 
     #[Groups(['api:list', 'api:single', 'snapshot'])]
     public ?string $country = null;
+
+    /**
+     * Raw `kostenstelle`, deliberately a string: coverage is thin and some hotels carry a
+     * space-separated list such as "84 85", so it is an opaque label, not a number.
+     */
+    #[Groups(['api:list', 'api:single', 'snapshot'])]
+    public ?string $costUnit = null;
+
+    /** {@see \App\BusProNet\Utility\CountryCodeMapper} — lowercase ISO-2, null when unmappable. */
+    #[Groups(['api:list', 'api:single', 'snapshot'])]
+    public ?string $countryCode = null;
 }

src/BusProNet/Utility/CountryCodeMapper.php

@@ -0,0 +1,120 @@
+<?php
+
+declare(strict_types=1);
+
+namespace App\BusProNet\Utility;
+
+use Psr\Log\LoggerInterface;
+
+/**
+ * Maps BusPro's `hotel/land` token onto a lowercase ISO 3166-1 alpha-2 code.
+ *
+ * BusPro has no country vocabulary: the field is free text that has accumulated three
+ * conventions side by side — ISO-2 (`CH`, `NL`), the European vehicle registration codes
+ * (`D`, `A`, `F`, `S`, `FL`) and ISO-3 (`AUT`, `PRT`, `HRV`), with the same country often
+ * under several tokens. The table below is therefore part of the API contract rather than
+ * a derivation: consumers render a flag from it, and a wrong entry mislabels a destination
+ * silently, which is worse than no flag at all. An unknown token maps to null — never to a
+ * guess, never to an empty string.
+ */
+class CountryCodeMapper
+{
+    /**
+     * Keys are upper-cased tokens, values lowercase ISO-2.
+     *
+     * Entries worth their explanation:
+     * - `IR` is Ireland, not Iran. Every row carrying it is Dublin or Carlow.
+     * - `SLO` is Slovakia here, not Slovenia: its only row is a Slovak company, and
+     *   Slovenia already arrives under its own `SI`.
+     * - `S`, `N`, `FL` are vehicle codes (Sweden, Norway, Liechtenstein), consistent with
+     *   the `D`/`A`/`F`/`I`/`E`/`B` in the same column.
+     * - `UK` is the colloquial form of the ISO `GB`.
+     *
+     * @var array<string, string>
+     */
+    private const MAP = [
+        'A' => 'at',
+        'AT' => 'at',
+        'AUT' => 'at',
+        'B' => 'be',
+        'CA' => 'ca',
+        'CH' => 'ch',
+        'CZ' => 'cz',
+        'D' => 'de',
+        'DEU' => 'de',
+        'DK' => 'dk',
+        'E' => 'es',
+        'ES' => 'es',
+        'F' => 'fr',
+        'FI' => 'fi',
+        'FL' => 'li',
+        'FR' => 'fr',
+        'GR' => 'gr',
+        'GRE' => 'gr',
+        'HR' => 'hr',
+        'HRV' => 'hr',
+        'HU' => 'hu',
+        'I' => 'it',
+        'IL' => 'il',
+        'IR' => 'ie',
+        'IS' => 'is',
+        'IT' => 'it',
+        'LT' => 'lt',
+        'LV' => 'lv',
+        'MT' => 'mt',
+        'N' => 'no',
+        'NL' => 'nl',
+        'PL' => 'pl',
+        'PRT' => 'pt',
+        'PT' => 'pt',
+        'RUS' => 'ru',
+        'S' => 'se',
+        'SI' => 'si',
+        'SLO' => 'sk',
+        'UK' => 'gb',
+        'USA' => 'us',
+        'ZA' => 'za',
+    ];
+
+    /**
+     * Tokens already reported in this process, so a master data file full of the same
+     * unknown token produces one log line rather than one per hotel.
+     *
+     * @var array<string, true>
+     */
+    private array $reported = [];
+
+    public function __construct(
+        private readonly LoggerInterface $logger,
+    ) {
+    }
+
+    /**
+     * @param string|null $token the raw `hotel/land` value
+     *
+     * @return string|null lowercase ISO-2, or null when absent, empty or unmappable
+     */
+    public function map(?string $token): ?string
+    {
+        if (null === $token) {
+            return null;
+        }
+
+        $normalized = strtoupper(trim($token));
+
+        if ('' === $normalized) {
+            return null;
+        }
+
+        if (isset(self::MAP[$normalized])) {
+            return self::MAP[$normalized];
+        }
+
+        if (false === isset($this->reported[$normalized])) {
+            $this->reported[$normalized] = true;
+            $this->logger->warning('Unmapped BusPro country token', ['token' => $token]);
+        }
+
+        return null;
+    }
+}

src/BusProNet/XmlLoader/HotelLoader.php

@@ -4,14 +4,25 @@ namespace App\BusProNet\XmlLoader;
 
 use App\BusProNet\Model\Hotel;
 use App\BusProNet\Model\Travel;
+use App\BusProNet\Utility\CountryCodeMapper;
 use App\BusProNet\XmlCrawlerFactory;
 use App\Exception\HotelNotFoundException;
+use League\Flysystem\FilesystemOperator;
 use Psr\Cache\InvalidArgumentException;
 use Symfony\Component\DomCrawler\Crawler;
+use Symfony\Contracts\Cache\CacheInterface;
 use Symfony\Contracts\Cache\ItemInterface;
 
 class HotelLoader extends AbstractLoader
 {
+    public function __construct(
+        private readonly CountryCodeMapper $countryCodeMapper,
+        CacheInterface $cache,
+        FilesystemOperator $xmlExport,
+    ) {
+        parent::__construct($cache, $xmlExport);
+    }
+
     /** @return array<int, Hotel> */
     public function loadAll(?string $filename = 'hotel.xml'): array
     {
@@ -87,6 +98,10 @@ class HotelLoader extends AbstractLoader
         $hotel->country = $this->getStringOrNullValue($node->filterXPath('//land'));
         $hotel->street = $this->getStringOrNullValue($node->filterXPath('//strasse'));
         $hotel->phone = $this->getStringOrNullValue($node->filterXPath('//telefon'));
+        // String, never int: see Hotel::$costUnit.
+        $hotel->costUnit = $this->getStringOrNullValue($node->filterXPath('//kostenstelle'));
+        // `country` itself stays the raw token — it is the join key into /api/countries[].token.
+        $hotel->countryCode = $this->countryCodeMapper->map($hotel->country);
 
         return $hotel;
     }