Question

我尝试将图片上传到Google云端硬盘以进行光学字符识别（OCR）。这是我的代码：

require_once('vendor/autoload.php');

// Initialize Google Client
$client_email = 'xxxxxx@yyyyy.iam.gserviceaccount.com';
$private_key = file_get_contents('key.p12');
$scopes = array(
    'https://www.googleapis.com/auth/drive.file'
);
$credentials = new Google_Auth_AssertionCredentials(
    $client_email,
    $scopes,
    $private_key
);
$client = new Google_Client();
$client->setAssertionCredentials($credentials);
if ($client->getAuth()->isAccessTokenExpired()) {
  $client->getAuth()->refreshTokenWithAssertion();
}

// Initialize Google Drive service
$service = new Google_Service_Drive($client);

// Upload File
$file = new Google_Service_Drive_DriveFile();
$file->setName('Test Image for OCR');
$file->setDescription('Test Image for OCR');
$file->setMimeType('image/jpeg');
try {
  $data = file_get_contents($filename);
  $createdFile = $service->files->create($file, array(
      'data' => $data,
      'mimeType' => 'image/jpeg',
  ));
  var_dump($createdFile);
  // ===========
  // So, what's next?
  // ===========
} catch(Exception $e) {
  echo 'Error occurred: ' . $e->getMessage();
}

以上代码运行时没有错误，$createdFile是Google_Service_Drive_DriveFile对象形式的有效资源。

问题：

我猜上传成功，因为create()函数没有返回错误。但是，我无法在我的Google云端硬盘中看到这些文件已上传。不应该将其上传到Google云端硬盘的根文件夹吗？
如何执行OCR？我可以从here读取有一个名为ocrLanguage的参数。我应该把它放在哪里，如何获得结果？

提前致谢。

更新 var_dump()结果如下：

object(Google_Service_Drive_DriveFile)#18 (55) {
  ["collection_key":protected]=>
  string(6) "spaces"
  ["internal_gapi_mappings":protected]=>
  array(0) {
  }
  ["appProperties"]=>
  NULL
  ["capabilitiesType":protected]=>
  string(42) "Google_Service_Drive_DriveFileCapabilities"
  ["capabilitiesDataType":protected]=>
  string(0) ""
  ["contentHintsType":protected]=>
  string(42) "Google_Service_Drive_DriveFileContentHints"
  ["contentHintsDataType":protected]=>
  string(0) ""
  ["createdTime"]=>
  NULL
  ["description"]=>
  NULL
  ["explicitlyTrashed"]=>
  NULL
  ["fileExtension"]=>
  NULL
  ["folderColorRgb"]=>
  NULL
  ["fullFileExtension"]=>
  NULL
  ["headRevisionId"]=>
  NULL
  ["iconLink"]=>
  NULL
  ["id"]=>
  string(28) "0B_XXXXX1yjq7dENaQWp4ckZoRk0"
  ["imageMediaMetadataType":protected]=>
  string(48) "Google_Service_Drive_DriveFileImageMediaMetadata"
  ["imageMediaMetadataDataType":protected]=>
  string(0) ""
  ["kind"]=>
  string(10) "drive#file"
  ["lastModifyingUserType":protected]=>
  string(25) "Google_Service_Drive_User"
  ["lastModifyingUserDataType":protected]=>
  string(0) ""
  ["md5Checksum"]=>
  NULL
  ["mimeType"]=>
  string(10) "image/jpeg"
  ["modifiedByMeTime"]=>
  NULL
  ["modifiedTime"]=>
  NULL
  ["name"]=>
  string(18) "Test Image for OCR"
  ["originalFilename"]=>
  NULL
  ["ownedByMe"]=>
  NULL
  ["ownersType":protected]=>
  string(25) "Google_Service_Drive_User"
  ["ownersDataType":protected]=>
  string(5) "array"
  ["parents"]=>
  NULL
  ["permissionsType":protected]=>
  string(31) "Google_Service_Drive_Permission"
  ["permissionsDataType":protected]=>
  string(5) "array"
  ["properties"]=>
  NULL
  ["quotaBytesUsed"]=>
  NULL
  ["shared"]=>
  NULL
  ["sharedWithMeTime"]=>
  NULL
  ["sharingUserType":protected]=>
  string(25) "Google_Service_Drive_User"
  ["sharingUserDataType":protected]=>
  string(0) ""
  ["size"]=>
  NULL
  ["spaces"]=>
  NULL
  ["starred"]=>
  NULL
  ["thumbnailLink"]=>
  NULL
  ["trashed"]=>
  NULL
  ["version"]=>
  NULL
  ["videoMediaMetadataType":protected]=>
  string(48) "Google_Service_Drive_DriveFileVideoMediaMetadata"
  ["videoMediaMetadataDataType":protected]=>
  string(0) ""
  ["viewedByMe"]=>
  NULL
  ["viewedByMeTime"]=>
  NULL
  ["viewersCanCopyContent"]=>
  NULL
  ["webContentLink"]=>
  NULL
  ["webViewLink"]=>
  NULL
  ["writersCanShare"]=>
  NULL
  ["modelData":protected]=>
  array(0) {
  }
  ["processed":protected]=>
  array(0) {
  }
}

该文件可以通过$service->files->get($file_id);获取，但在我的Google云端硬盘中不可见。返回的文件资源对象也不包含任何有用的内容。

Answer 1

我刚刚在V3中找到了通往OCR的方法。

上传图片
使用mimeType ＆＃34; application / vnd.google-apps.document＆＃34;
使用mimeType ＆＃34; text / plain＆＃34;

P.S。似乎第2步赢得了＆＃34; appDataFolder＆＃34;。

UserCredential credential = null;
try
{
    credential = await GoogleWebAuthorizationBroker.AuthorizeAsync(
        new Uri("ms-appx:///Assets/client_secret.json"), 
        new[] { DriveService.Scope.DriveFile }, "user", CancellationToken.None);
}
catch (AggregateException ex)
{
    Debug.Write("Credential failed, " + ex.Message);
}

// Create Drive API service.
var service = new DriveService(new BaseClientService.Initializer()
{
    HttpClientInitializer = credential,
    ApplicationName = "TestApp",
});

// Create folder
var folderMetadata = new Google.Apis.Drive.v3.Data.File();
folderMetadata.Name = "NewFolder";
folderMetadata.MimeType = "application/vnd.google-apps.folder";
var request = service.Files.Create(folderMetadata);
request.Fields = "id";
var folder = request.Execute();
Debug.WriteLine("Folder ID: " + folder.Id);

// Upload the image file
var fileMetadata = new Google.Apis.Drive.v3.Data.File();
fileMetadata.Name = inputFile.Name;
fileMetadata.Parents = new List<string> { folder.Id };
FilesResource.CreateMediaUpload requestUpload;
using (var stream = new System.IO.FileStream(inputFile.Path, System.IO.FileMode.Open))
{
    requestUpload = service.Files.Create(fileMetadata, stream, "image/jpeg");
    requestUpload.Fields = "id";
    requestUpload.Upload();
}
var imgFile = requestUpload.ResponseBody;
Debug.WriteLine("File ID: " + imgFile.Id);

// Copy image and paste as document
var textMetadata = new Google.Apis.Drive.v3.Data.File();
textMetadata.Name = inputFile.Name;
textMetadata.Parents = new List<string> { folderId };
textMetadata.MimeType = "application/vnd.google-apps.document";
FilesResource.CopyRequest requestCopy = service.Files.Copy(textMetadata, imgFile.Id);
requestCopy.Fields = "id";
requestCopy.OcrLanguage = "zh";
var textFile = requestCopy.Execute();

// Now we export document as plain text
FilesResource.ExportRequest requestExport = service.Files.Export(textFile.Id, "text/plain");
string output = requestExport.Execute();

Answer 2

服务帐户不是你更像虚拟用户。它有自己的驱动器帐户。

如果您想上传到您的个人帐户。获取服务帐户电子邮件地址并将其分享到个人驱动器帐户中的目录。就像您希望与其他任何用户共享目录或文件一样。

然后你需要找出目录ID，我发现这样做的方法是让服务帐户执行files.list以获取它现在可以访问的所有内容的列表。找到目录ID或父ID后您可以将上面的代码更改为

return_type func_name(type param_name1, type param_name2, …,type param_nameN);

我认为你使用V3 api的代码看起来我没有时间玩它。 'data' => $data, 'mimeType' => 'image/jpeg', 'parents' => 'the directory id'＆lt; - 这是一个有根据的猜测。如果它不起作用，请告诉我，我会稍微了解一下如何将父母传递给v3。

选项nr 2：

另一个选项是服务帐户与您共享其文件夹，然后您将有权访问其驱动器帐户，并且您将能够在您的网络版Drive中看到该文件夹。再次搜索权限我认为你正在使用V3我还没有看过它。差异在于存储数据的位置以及存储计数的对象。

Answer 3

请勿使用服务帐户。如果您要上传到自己的帐户，则只需为您的帐户获取适当的访问令牌即可。使用共享文件夹的中间帐户非常难看（imho）。

将图片上传到Google云端硬盘以进行OCR

3 个答案: