Cache OCR transcriptions

1 hour by default

Bug: T282837

.env

@@ -31,3 +31,5 @@ APP_LOG_SUBJECT="[Wikimedia OCR]"
 
 # Comma-separated list of the host names (without protocols) of where images are stored.
 APP_IMAGE_HOSTS=upload.wikimedia.org,upload.wikimedia.beta.wmflabs.org
+
+APP_CACHE_TTL=3600 # 1 hour

composer.json

@@ -11,6 +11,7 @@
         "ext-json": "*",
         "google/cloud-vision": "^1.3",
         "sensio/framework-extra-bundle": "^6.1",
+        "symfony/cache": "5.2.*",
         "symfony/console": "5.2.*",
         "symfony/dotenv": "5.2.*",
         "symfony/flex": "^1.3.1",

config/services.yaml

@@ -4,6 +4,7 @@
 # Put parameters here that don't need to change on each machine where the app is deployed
 # https://symfony.com/doc/current/best_practices/configuration.html#application-related-configuration
 parameters:
+    cache_ttl: '%env(APP_CACHE_TTL)%'
 
 services:
     # default configuration for services in *this* file

src/Controller/OcrController.php

@@ -13,6 +13,8 @@
 use Symfony\Component\HttpFoundation\RequestStack;
 use Symfony\Component\HttpFoundation\Response;
 use Symfony\Component\Routing\Annotation\Route;
+use Symfony\Contracts\Cache\CacheInterface;
+use Symfony\Contracts\Cache\ItemInterface;
 
 class OcrController extends AbstractController
 {
@@ -22,6 +24,9 @@ class OcrController extends AbstractController
     /** @var TesseractEngine|GoogleCloudVisionEngine */
     protected $engine;
 
+    /** @var CacheInterface */
+    protected $cache;
+
     /**
      * The output params for the view or API response.
      * This also serves as where you define the defaults.
@@ -43,11 +48,17 @@ class OcrController extends AbstractController
      * @param RequestStack $requestStack
      * @param Intuition $intuition
      * @param EngineFactory $engineFactory
+     * @param CacheInterface $cache
      */
-    public function __construct(RequestStack $requestStack, Intuition $intuition, EngineFactory $engineFactory)
-    {
+    public function __construct(
+        RequestStack $requestStack,
+        Intuition $intuition,
+        EngineFactory $engineFactory,
+        CacheInterface $cache
+    ) {
         // Dependencies.
         $this->intuition = $intuition;
+        $this->cache = $cache;
 
         $request = $requestStack->getCurrentRequest();
 
@@ -118,7 +129,7 @@ public function homeAction(): Response
         static::$params['image_hosts'] = $this->intuition->listToText(static::$params['image_hosts']);
 
         if ($this->imageUrl) {
-            static::$params['text'] = $this->engine->getText($this->imageUrl, static::$params['langs']);
+            static::$params['text'] = $this->getText();
         }
 
         return $this->render('output.html.twig', static::$params);
@@ -132,7 +143,7 @@ public function homeAction(): Response
     public function apiAction(): JsonResponse
     {
         return $this->getApiResponse(array_merge(static::$params, [
-            'text' => $this->engine->getText($this->imageUrl, static::$params['langs']),
+            'text' => $this->getText(),
         ]));
     }
 
@@ -165,4 +176,27 @@ private function getApiResponse(array $params): JsonResponse
         $response->setData($params);
         return $response;
     }
+
+    /**
+     * Get and cache the transcription based on options set in static::$params.
+     * @return string
+     */
+    private function getText(): string
+    {
+        $cacheKey = md5(implode(
+            '|',
+            [
+                $this->imageUrl,
+                static::$params['engine'],
+                implode('|', static::$params['langs']),
+                static::$params['psm'],
+                static::$params['oem'],
+            ]
+        ));
+
+        return $this->cache->get($cacheKey, function (ItemInterface $item) {
+            $item->expiresAfter((int)$this->getParameter('cache_ttl'));
+            return $this->engine->getText($this->imageUrl, static::$params['langs']);
+        });
+    }
 }

tests/Controller/OcrControllerTest.php

@@ -9,6 +9,7 @@
 use App\Engine\TesseractEngine;
 use Krinkle\Intuition\Intuition;
 use PHPUnit\Framework\TestCase;
+use Symfony\Component\Cache\Adapter\FilesystemAdapter;
 use Symfony\Component\HttpClient\MockHttpClient;
 use Symfony\Component\HttpFoundation\Request;
 use Symfony\Component\HttpFoundation\RequestStack;
@@ -32,7 +33,8 @@ public function testGetLang(array $getParams, array $expectedLangs): void
         $controller = new OcrController(
             $requestStack,
             $intuition,
-            new EngineFactory($gcv, new TesseractEngine(new MockHttpClient(), $intuition, new TesseractOCR()))
+            new EngineFactory($gcv, new TesseractEngine(new MockHttpClient(), $intuition, new TesseractOCR())),
+            new FilesystemAdapter()
         );
         $this->assertSame($expectedLangs, $controller->getLangs($request));
     }