Google Docs API: Client Login with PHP and Curl

A few days ago I started looking deeper into the Google Code APIs and threw a few experiments using the Google Documents List Data API. Unfortunately, the only library they have for the third version of their protocol is written in Java. There is a PHP wrapper for the first version of the protocol, but it totally depends on the Zend Framework. Here’s a little code snippet for logging into a Google Docs account (writely) using ClientLogin with Curl and PHP. The Auth string is stored into the $auth variable for later use.

// Construct an HTTP POST request
$clientlogin_url = "https://www.google.com/accounts/ClientLogin";
$clientlogin_post = array(
    "accountType" => "HOSTED_OR_GOOGLE",
    "Email" => "yourgoogle@email.com",
    "Passwd" => "yourgooglepassword",
    "service" => "writely",
    "source" => "your application name"
);

// Initialize the curl object
$curl = curl_init($clientlogin_url);

// Set some options (some for SHTTP)
curl_setopt($curl, CURLOPT_POST, true);
curl_setopt($curl, CURLOPT_POSTFIELDS, $clientlogin_post);
curl_setopt($curl, CURLOPT_HTTPAUTH, CURLAUTH_ANY);
curl_setopt($curl, CURLOPT_SSL_VERIFYPEER, false);
curl_setopt($curl, CURLOPT_RETURNTRANSFER, 1);

// Execute
$response = curl_exec($curl);

// Get the Auth string and save it
preg_match("/Auth=([a-z0-9_-]+)/i", $response, $matches);
$auth = $matches[1];

echo "The auth string is: " . $auth;

ClientLogin assumes you login once and use the auth string for all on-going requests (more info: ClientLogin – Google Code). Here’s a simple example of how to retrieve some data from a Google Docs account, show filename, author and file type. Make sure the simplexml php extension is installed and activated.

// Include the Auth string in the headers
// Together with the API version being used
$headers = array(
    "Authorization: GoogleLogin auth=" . $auth,
    "GData-Version: 3.0",
);

// Make the request
curl_setopt($curl, CURLOPT_URL, "http://docs.google.com/feeds/default/private/full");
curl_setopt($curl, CURLOPT_HTTPHEADER, $headers);
curl_setopt($curl, CURLOPT_POST, false);

$response = curl_exec($curl);
curl_close($curl);

// Parse the response
$response = simplexml_load_string($response);

// Output data
foreach($response->entry as $file) {
	echo "File: " . $file->title . "<br />";
	echo "Type: " . $file->content["type"] . "<br />";
	echo "Author: " . $file->author->name . "<br /><br />";
}

It seems though that Google Docs is purely for documents (spreadsheets and presentations) but not for actual files, as a JPEG I uploaded turned into a text/html type. It opens up in the text editor when viewing in Google Docs, and the nearest-to JPEG export format is PDF (just like all the other documents). So I guess there’s no way of reading and manipulating the raw image as if I would read and manipulate a file.

I came across a cool website which I believe the Googlers have made – Data Liberation:

Users should be able to control the data they store in any of Google’s products. Our team’s goal is to make it easier for them to move data in and out.

I’m working on a little project that involves image (and file) hosting, so the products I’m interested in are Google Docs and Picasa Web Albums. It seems that none of the two make it possible to store and work with files as if it was a flash disk or an FTP server. Google Docs is the closest one but there’s a huge limit on what kind of files you can upload and store. Picasa on the other hand is good for image storing, but the only way to export a set of images is to use their desktop software and Picasa Web, and of course no actual control over the files (edit, delete) via their Picasa Web API. If only it were a little bit closer to Amazon S3 …

Oh and you should probably use OAuth or AuthSub when working with Web Applications and Google APIs but anyways, this was just a quick example.

Twitter API: Picking the Right Source

I’m sure you noticed that a few weeks ago Twitter changed the source that came unsigned via the API from web to API which could basically reveal any robot that is trying to act human, right? Well if you look at the statuses/update method in the Twitter API documentation they don’t say anything about the source parameter. Strange, right?

Well I read something about it on some forums and the Twitter API development talk Google Group had a discussion about this I believe, but most people still think that it’s the X-Twitter- headers that determine the source, but it’s not. It’s way too simple and it took me a few hours to figure it out using an HTTP sniffer on TweetDeck, Seesmic Desktop and the others.

Turns out it’s the source parameter that is passed via POST together with the status text during the statuses/update call. So usually you would do something like:

$postargs = array("status" => "I'm tweeting via API!");
curl_setopt ($ch, CURLOPT_POSTFIELDS, $postargs);

Now, how about Seesmic Desktop?

$postargs = array(
    "status" => "I'm tweeting via Seesmic Desktop!",
    "source" => "seesmicdesktop"
);
curl_setopt ($ch, CURLOPT_POSTFIELDS, $postargs);

Or TweetDeck perhaps?

$postargs = array(
    "status" => "I'm tweeting via TweetDeck!",
    "source" => "tweetdeck"
);
curl_setopt ($ch, CURLOPT_POSTFIELDS, $postargs);

Pretty cool, huh? Here’s a brief list of sources to pick from:

And I do believe that it’s the same way for all the other clients. Lowercase and no spaces. It is funny though watching a robot tweet 10 tweets per minute via TwitterFon ;)