Tesseract 圖片識別

代碼倉庫 Hyperf tesseract-ocr-for-php

建立項目

composer create hyperf/biz-skeleton tesseract-demo
複製代碼

安裝擴展

composer require thiagoalessio/tesseract_ocr
複製代碼

編寫接口

編寫單元測試

首先，咱們編寫接口單元測試。當沒有傳 url 時，錯誤碼返回 1000，當傳入圖片地址後，返回對應的識別信息。

<?php

declare(strict_types=1);
/** * This file is part of Hyperf. * * @link https://www.hyperf.io * @document https://doc.hyperf.io * @contact group@hyperf.io * @license https://github.com/hyperf-cloud/hyperf/blob/master/LICENSE */

namespace HyperfTest\Cases;

use HyperfTest\HttpTestCase;

/** * @internal * @coversNothing */
class ImageTest extends HttpTestCase {
    public function testImageReadFailed() {
        $res = $this->json('/image/read');
        $this->assertSame(1000, $res['code']);
    }

    public function testImageRead() {
        $res = $this->json('/image/read', [
            'url' => 'https://raw.githubusercontent.com/thiagoalessio/tesseract-ocr-for-php/master/tests/EndToEnd/images/8055.png',
        ]);

        $this->assertSame(0, $res['code']);
        $this->assertSame('8055', $res['data']);
    }
}

複製代碼

添加 Task

由於咱們要使用的 thiagoalessio/tesseract_ocr 庫不肯定是否能夠進行協程調度，因此咱們把具體邏輯放到 Task 中執行。

composer require hyperf/task
複製代碼

改造咱們的 server.php

<?php

declare(strict_types=1);

use Hyperf\Server\SwooleEvent;

return [
    // 這裏省略了其它不相關的配置項
    'settings' => [
        // Task Worker 數量，根據您的服務器配置而配置適當的數量
        'task_worker_num' => 8,
        // 由於 `Task` 主要處理沒法協程化的方法，因此這裏推薦設爲 `false`，避免協程下出現數據混淆的狀況
        'task_enable_coroutine' => false,
    ],
    'callbacks' => [
        // Task callbacks
        SwooleEvent::ON_TASK => [Hyperf\Framework\Bootstrap\TaskCallback::class, 'onTask'],
        SwooleEvent::ON_FINISH => [Hyperf\Framework\Bootstrap\FinishCallback::class, 'onFinish'],
    ],
];
複製代碼

編寫實現邏輯

實現邏輯，很是簡單，咱們把網圖保存到本地，而後經過 tesseract 來讀取便可。

<?php

declare(strict_types=1);
/** * This file is part of Hyperf. * * @link https://www.hyperf.io * @document https://doc.hyperf.io * @contact group@hyperf.io * @license https://github.com/hyperf-cloud/hyperf/blob/master/LICENSE */

namespace App\Service;

use Hyperf\Guzzle\ClientFactory;
use Hyperf\Task\Annotation\Task;
use thiagoalessio\TesseractOCR\TesseractOCR;

class ImageService {
    public function read(string $url) {
        $path = $this->save($url);

        return $this->tesseract($path);
    }

    /** * @Task */
    public function tesseract(string $path) {
        return (new TesseractOCR($path))->run();
    }

    protected function save(string $url): string {
        $client = di()->get(ClientFactory::class)->create();
        $content = $client->get($url)->getBody()->getContents();

        $path = BASE_PATH . '/runtime/' . uniqid();
        file_put_contents($path, $content);

        return $path;
    }
}

複製代碼

控制器代碼以下

<?php

declare(strict_types=1);
/** * This file is part of Hyperf. * * @link https://www.hyperf.io * @document https://doc.hyperf.io * @contact group@hyperf.io * @license https://github.com/hyperf-cloud/hyperf/blob/master/LICENSE */

namespace App\Controller;

use App\Constants\ErrorCode;
use App\Exception\BusinessException;
use App\Service\ImageService;
use Hyperf\Di\Annotation\Inject;
use Hyperf\HttpServer\Annotation\AutoController;

/** * @AutoController(prefix="/image") */
class ImageController extends Controller {
    /** * @Inject * @var ImageService */
    protected $service;

    public function read() {
        $url = $this->request->input('url');

        if (empty($url)) {
            throw new BusinessException(ErrorCode::PARAMS_INVALID);
        }

        $result = $this->service->read($url);

        return $this->response->success($result);
    }
}

複製代碼

執行單元測試

$ composer test
> co-phpunit -c phpunit.xml --colors=always
Scanning ...
Scan completed.
[DEBUG] Event Hyperf\Framework\Event\BootApplication handled by Hyperf\Di\Listener\BootApplicationListener listener.
[DEBUG] Event Hyperf\Framework\Event\BootApplication handled by Hyperf\Config\Listener\RegisterPropertyHandlerListener listener.
[DEBUG] Event Hyperf\Framework\Event\BootApplication handled by Hyperf\Paginator\Listener\PageResolverListener listener.
PHPUnit 7.5.14 by Sebastian Bergmann and contributors.

...                                                                 3 / 3 (100%)

Time: 1.11 seconds, Memory: 42.00 MB

OK (3 tests, 13 assertions)

複製代碼

固然，也能夠經過 CURL 訪問，查看結果

$ curl http://127.0.0.1:9501/image/read -H 'Content-Type:application/json' -d '{"url":"https://raw.githubusercontent.com/thiagoalessio/tesseract-ocr-for-php/master/tests/EndToEnd/images/8055.png"}'
{"code":0,"data":"8055"}%
複製代碼

修改 Docker

由於默認的 Docker環境 沒有 tesseract，因此咱們修改一下 Dockerfile

# Default Dockerfile
#
# @link https://www.hyperf.io
# @document https://doc.hyperf.io
# @contact group@hyperf.io
# @license https://github.com/hyperf-cloud/hyperf/blob/master/LICENSE

FROM hyperf/hyperf:7.2-alpine-cli
LABEL maintainer="Hyperf Developers <group@hyperf.io>" version="1.0" license="MIT" 
##
# ---------- env settings ----------
##
# --build-arg timezone=Asia/Shanghai
ARG timezone

ENV TIMEZONE=${timezone:-"Asia/Shanghai"} \
    COMPOSER_VERSION=1.8.6 \
    APP_ENV=prod

# update
RUN set -ex \ && apk update \ # 安裝 tesseract-ocr     && apk add tesseract-ocr \ # install composer     && cd /tmp \
    && wget https://github.com/composer/composer/releases/download/${COMPOSER_VERSION}/composer.phar \
    && chmod u+x composer.phar \
    && mv composer.phar /usr/local/bin/composer \
    # show php version and extensions
    && php -v \
    && php -m \
    # ---------- some config ----------
    && cd /etc/php7 \
    # - config PHP
    && { \
        echo "upload_max_filesize=100M"; \
        echo "post_max_size=108M"; \
        echo "memory_limit=1024M"; \
        echo "date.timezone=${TIMEZONE}"; \
    } | tee conf.d/99-overrides.ini \
    # - config timezone
    && ln -sf /usr/share/zoneinfo/${TIMEZONE} /etc/localtime \
    && echo "${TIMEZONE}" > /etc/timezone \
    # ---------- clear works ----------
    && rm -rf /var/cache/apk/* /tmp/* /usr/share/man \
    && echo -e "\033[42;37m Build Completed :).\033[0m\n"

WORKDIR /opt/www 
COPY . /opt/www 
RUN composer install --no-dev \ && composer dump-autoload -o \ && php /opt/www/bin/hyperf.php di:init-proxy 
EXPOSE 9501

ENTRYPOINT ["php", "/opt/www/bin/hyperf.php", "start"] 
複製代碼

構造鏡像

docker build -t tesseract-demo .
複製代碼

啓動鏡像

docker run --rm -p 9501:9501 -d --name tesseract-demo tesseract-demo
複製代碼

測試

$ curl http://127.0.0.1:9501/image/read -H 'Content-Type:application/json' -d '{"url":"https://raw.githubusercontent.com/thiagoalessio/tesseract-ocr-for-php/master/tests/EndToEnd/images/8055.png"}'
{"code":0,"data":"8055"}%
複製代碼