Reputation: 1929
I have some code working that uploads a video file up to YouTube:
$yt = new Zend_Gdata_YouTube($httpClient);
// create a new VideoEntry object
$myVideoEntry = new Zend_Gdata_YouTube_VideoEntry();
// create a new Zend_Gdata_App_MediaFileSource object
$filesource = $yt->newMediaFileSource('file.mov');
$filesource->setContentType('video/quicktime');
// set slug header
$filesource->setSlug('file.mov');
I have videos in S3 and I want to upload them to YouTube. The video in our S3 account is public, so i can use a command like wget. Should I run a command that wgets the video file and downloads it locally before I run this script (shell_exec("wget ".$s3videoURL)
)?
Or should I try to enter the MediaFileSource as the URL of the S3 file itself?
Mainly, I just need stability (not a solution subject to frequent time-outs); speed and local storage isn't really important (I can locally delete video file once its been uploaded).
What would be the best way to go about this?
Thanks!
Update: I should probably mention that this script is going to be uploading about 5 videos to YouTube per execution.
Upvotes: 6
Views: 5549
Reputation: 2910
I had to make some changes to @previous_developer 's answer to make it work with Youtube Data API V3 (Please upvote him as I could not find any working code except his).
$streamName = 's3://BUCKET-NAME/VIDEO.mp4';
/**
Since I have been using Yii 2. Use the AWS
SDK directly instead.
*/
$aws = Yii::$app->awssdk->getAwsSdk();
$s3client = $aws->createS3();
$s3client->registerStreamWrapper();
$service = new \Google_Service_YouTube($client);
$snippet = new \Google_Service_YouTube_VideoSnippet();
$snippet->setTitle("Test title");
$snippet->setDescription("Test descrition");
$snippet->setTags(array("tag1","tag2"));
$snippet->setCategoryId("22");
$status = new \Google_Service_YouTube_VideoStatus();
$status->privacyStatus = "public";
$video = new \Google_Service_YouTube_Video();
$video->setSnippet($snippet);
$video->setStatus($status);
$client->setDefer(true);
$insertResponse = $service->videos->insert("status,snippet", $video);
$media = new MediaFileUpload(
$client,
$insertResponse,
'video/*',
null,
true,
false
);
$filesize = filesize($streamName); // use it as a reguler file.
$media->setFileSize($filesize);
$chunkSizeBytes = 2 * 1024 * 1024; // 2 mb
$uploadStatus = false;
$handle = fopen($streamName, "r");
$totalSend = 0;
$totalReceived = 0;
$chunkBuffer = '';
while (!$uploadStatus && !feof($handle)) {
$chunk = fread($handle, $chunkSizeBytes);
$chunkBuffer .= $chunk;
$chunkBufferSize = strlen($chunkBuffer);
if($chunkBufferSize > $chunkSizeBytes) {
$fullChunk = substr($chunkBuffer, 0, $chunkSizeBytes);
$leapChunk = substr($chunkBuffer, $chunkSizeBytes);
$uploadStatus = $media->nextChunk($fullChunk);
$totalSend += strlen($fullChunk);
$chunkBuffer = $leapChunk;
echo PHP_EOL.'Status: '.($totalReceived).' / '.$filesize.' (%'.(($totalReceived / $filesize) * 100).')'.PHP_EOL;
}
$totalReceived += strlen($chunk);
}
$extraChunkLen = strlen($chunkBuffer);
$uploadStatus = $media->nextChunk($chunkBuffer);
$totalSend += strlen($chunkBuffer);
fclose($handle);
// If you want to make other calls after the file upload, set setDefer back to false
$client->setDefer(false);
Upvotes: 1
Reputation: 143
$chunkSizeBytes = 2 * 1024 * 1024; // 2 mb
$s3client = $this->c_aws->getS3Client();
$s3client->registerStreamWrapper();
try {
$client = new \Google_Client();
$client->setAccessType("offline");
$client->setApprovalPrompt('force');
$client->setClientId(GOOGLE_CLIENT_ID);
$client->setClientSecret(GOOGLE_CLIENT_SECRET);
$token = $client->fetchAccessTokenWithRefreshToken(GOOGLE_REFRESH_TOKEN);
$client->setAccessToken($token);
$youtube = new \Google_Service_YouTube($client);
// Create a snippet with title, description, tags and category ID
// Create an asset resource and set its snippet metadata and type.
// This example sets the video's title, description, keyword tags, and
// video category.
$snippet = new \Google_Service_YouTube_VideoSnippet();
$snippet->setTitle($title);
$snippet->setDescription($summary);
$snippet->setTags(explode(',', $keywords));
// Numeric video category. See
// https://developers.google.com/youtube/v3/docs/videoCategories/list
// $snippet->setCategoryId("22");
// Set the video's status to "public". Valid statuses are "public",
// "private" and "unlisted".
$status = new \Google_Service_YouTube_VideoStatus();
$status->privacyStatus = "public";
// Associate the snippet and status objects with a new video resource.
$video = new \Google_Service_YouTube_Video();
$video->setSnippet($snippet);
$video->setStatus($status);
// Setting the defer flag to true tells the client to return a request which can be called
// with ->execute(); instead of making the API call immediately.
$client->setDefer(true);
$insertRequest = $youtube->videos->insert("status,snippet", $video);
$media = new \Google_Http_MediaFileUpload(
$client,
$insertRequest,
'video/*',
null,
true,
$chunkSizeBytes
);
$result = $this->c_aws->getAwsFile($aws_file_path);
$media->setFileSize($result['ContentLength']);
$uploadStatus = false;
// Seek to the beginning of the stream
$result['Body']->rewind();
// Read the body off of the underlying stream in chunks
while (!$uploadStatus && $data = $result['Body']->read($chunkSizeBytes)) {
$uploadStatus = $media->nextChunk($data);
}
$client->setDefer(false);
if ($uploadStatus->status['uploadStatus'] == 'uploaded') {
// Actions to perform for a successful upload
$uploaded_video_id = $uploadStatus['id'];
return ($uploadStatus['id']);
}
}catch (\Google_Service_Exception $exception){
return '';
print_r($exception);
}
Upvotes: -1
Reputation: 11028
This is an old question but i believe i have a better answer.
You don't have to write video to HDD and you can't keep the whole thing in RAM (I assume it is a big file).
You can use PHP AWS SDK and Google Client libraries to buffer file from S3 and send it to YouTube on the fly. Use registerStreamWrapper method to register S3 as file system and use resumable uploads from YouTube API. Then all you have to do is reading chunks from S3 with fread and sending them to YouTube. This way you can even limit the RAM usage.
I assume you created the video object ($video in code) from Google_Video class. This is a complete code.
<?php
require_once 'path/to/libraries/aws/vendor/autoload.php';
require_once 'path/to/libraries/google-client-lib/autoload.php';
use Aws\S3\S3Client;
$chunkSizeBytes = 2 * 1024 * 1024; // 2 mb
$streamName = 's3://bucketname/video.mp4';
$s3client = S3Client::factory(array(
'key' => S3_ACCESS_KEY,
'secret' => S3_SECRET_KEY,
'region' => 'eu-west-1' // if you need to set.
));
$s3client->registerStreamWrapper();
$client = new Google_Client();
$client->setClientId(YOUTUBE_CLIENT_ID);
$client->setClientSecret(YOUTUBE_CLIENT_SECRET);
$client->setAccessToken(YOUTUBE_TOKEN);
$youtube = new Google_YoutubeService($client);
$media = new Google_MediaFileUpload('video/*', null, true, $chunkSizeBytes);
$filesize = filesize($streamName); // use it as a reguler file.
$media->setFileSize($filesize);
$insertResponse = $youtube->videos->insert("status,snippet", $video, array('mediaUpload' => $media));
$uploadStatus = false;
$handle = fopen($streamName, "r");
$totalReceived = 0;
$chunkBuffer = '';
while (!$uploadStatus && !feof($handle)) {
$chunk = fread($handle, $chunkSizeBytes);
$chunkBuffer .= $chunk;
$chunkBufferSize = strlen($chunkBuffer);
if($chunkBufferSize > $chunkSizeBytes) {
$fullChunk = substr($chunkBuffer, 0, $chunkSizeBytes);
$leapChunk = substr($chunkBuffer, $chunkSizeBytes);
$uploadStatus = $media->nextChunk($insertResponse, $fullChunk);
$totalSend += strlen($fullChunk);
$chunkBuffer = $leapChunk;
echo PHP_EOL.'Status: '.($totalReceived).' / '.$filesize.' (%'.(($totalReceived / $filesize) * 100).')'.PHP_EOL;
}
$totalReceived += strlen($chunk);
}
$extraChunkLen = strlen($chunkBuffer);
$uploadStatus = $media->nextChunk($insertResponse, $chunkBuffer);
$totalSend += strlen($chunkBuffer);
fclose($handle);
Upvotes: 9
Reputation: 16705
The "MediaFileSource" must be a real file. It won't take a URL, so you will need to copy the videos to your server from S3, before sending them to YouTube.
You can probably get away with the "shell_exec" if your usage is light, but for a variety of reasons its probably better to use either the Zend S3 Service, or cURL to pull files from S3.
Upvotes: 2