i was away from Ruby for a while...now, i need to recreate a php api connection (curl) on ruby, to add it to my rails app, but all im getting is a Net::HTTPMovedPermanently message
here is the original php version:
$url = "https://marketplace.walmartapis.com/v3/token";
$uniqid = uniqid();
$authorization_key = base64_encode($client_id.":".$client_secret);
$code = "";
$ch = curl_init();
$options = array(
"WM_SVC.NAME: Walmart Marketplace",
"WM_QOS.CORRELATION_ID: $uniqid",
"Authorization: Basic $authorization_key",
"Accept: application/json",
"Content-Type: application/x-www-form-urlencoded",
),
);
curl_setopt_array($ch,$options);
$response = curl_exec($ch);
$code = curl_getinfo($ch,CURLINFO_HTTP_CODE);
curl_close($ch);
if($code == 201 || $code == 200)
{
$token = json_decode($response,true);
return $token['access_token'];
}
}
And this is my Ruby version so far:
require 'net/http'
require 'base64'
client_id = "id"
client_secret = "secret"
url = "https://marketplace.walmartapis.com/v3/token/"
uniqid = "1234567890abc"
uri = URI(url)
request = Net::HTTP::Post.new(uri)
request['WM_SVC.NAME'] = 'Walmart Marketplace'
request['WM_QOS.CORRELATION_ID'] = uniqid
request.basic_auth(client_id, client_secret)
request['Accept'] = 'application/json'
request['Content-Type'] = 'application/x-www-form-urlencoded'
http = Net::HTTP.new(uri.host)
response = http.request(request)
puts response
What im missing?
Update...
i've changed some things based on curl-to-ruby web, and is at follows...now im getting a 400 error
require 'net/http'
require 'uri'
client_id = "foo"
client_secret = "bar"
url = "https://marketplace.walmartapis.com/v3/token"
uniqid = "1234567890abc"
uri = URI.parse(url)
request = Net::HTTP::Post.new(uri)
request["WM_SVC.NAME"] = "Walmart Marketplace"
request["WM_QOS.CORRELATION_ID"] = uniqid
request.basic_auth(client_id, client_secret)
request["Accept"] = "application/json"
request.content_type = "application/x-www-form-urlencoded"
req_options = {
use_ssl: uri.scheme == "https",
}
response = Net::HTTP.start(uri.hostname, req_options) do |http|
http.request(request)
end
puts response.code
puts response.body
i created a new question to clear out whats working so far, and the missing fields comparing to php
It looks like the resource you are trying to access has been moved. The HTTPMovedPermanently reponse is a form of HTTPRedirection. Your code should make another request to the new location.
This is the example directly from the Net::HTTP documentation:
def fetch(uri_str, limit = 10)
# You should choose a better exception.
raise ArgumentError, 'too many HTTP redirects' if limit == 0
response = Net::HTTP.get_response(URI(uri_str))
case response
when Net::HTTPSuccess then
response
when Net::HTTPRedirection then
location = response['location']
warn "redirected to #{location}"
fetch(location, limit - 1)
else
response.value
end
end
I've made a simple script that posts images on tumblr.
everything is fine, but I've noticed some performance issues right after I've changed the host provider (my new host is limited and cheaper).
now, after debugging the script and after contacting the tumblr api helpdesk, I'm stuck on a problem:
there are 3 functions:
function oauth_gen($method, $url, $iparams, &$headers) {
$iparams['oauth_consumer_key'] = CONSUMER_KEY;
$iparams['oauth_nonce'] = strval(time());
$iparams['oauth_signature_method'] = 'HMAC-SHA1';
$iparams['oauth_timestamp'] = strval(time());
$iparams['oauth_token'] = OAUTH_TOKEN;
$iparams['oauth_version'] = '1.0';
$iparams['oauth_signature'] = oauth_sig($method, $url, $iparams);
$oauth_header = array();
foreach($iparams as $key => $value) {
if (strpos($key, "oauth") !== false) {
$oauth_header []= $key ."=".$value;
}
}
$str = print_r($iparams, true);
file_put_contents('data1-1.txt', $str);
$oauth_header = "OAuth ". implode(",", $oauth_header);
$headers["Authorization"] = $oauth_header;
}
function oauth_sig($method, $uri, $params) {
$parts []= $method;
$parts []= rawurlencode($uri);
$iparams = array();
ksort($params);
foreach($params as $key => $data) {
if(is_array($data)) {
$count = 0;
foreach($data as $val) {
$n = $key . "[". $count . "]";
$iparams []= $n . "=" . rawurlencode($val);
//$iparams []= $n . "=" . $val;
$count++;
}
} else {
$iparams[]= rawurlencode($key) . "=" .rawurlencode($data);
}
}
//debug($iparams,"iparams");
$str = print_r($iparams, true);
file_put_contents('data-1.txt', $str);
//$size = filesize('data.txt');
$parts []= rawurlencode(implode("&", $iparams));
//debug($parts,"parts");
//die();
$sig = implode("&", $parts);
return base64_encode(hash_hmac('sha1', $sig, CONSUMER_SECRET."&". OAUTH_SECRET, true));
}
these 2 functions above comes from an online functional example, they have always worked fine.
this is the function I use to call the APIs and the oAuth:
function posta_array($files,$queue,$tags,$caption,$link,$blog){
$datArr = array();
$photoset_layout = "";
foreach ($files as $sing_file){
$dataArr [] = file_get_contents($sing_file);
$photoset_layout .= "1";
}
$headers = array("Host" => "http://api.tumblr.com/", "Content-type" => "application/x-www-form-urlencoded", "Expect" => "");
$params = array(
"data" => $dataArr,
"type" => "photo",
"state" => $queue,
"tags"=>$tags,
"caption"=>$caption,
"photoset_layout" => $photoset_layout,
"link"=>str_replace("_","",$link)
);
debug($headers,"head");
oauth_gen("POST", "http://api.tumblr.com/v2/blog/$blog/post", $params, $headers);
debug($headers,"head 2");
$ch = curl_init();
curl_setopt($ch, CURLOPT_USERAGENT, "Tumblr v1.0");
curl_setopt($ch, CURLOPT_URL, "http://api.tumblr.com/v2/blog/$blog/post");
curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1 );
curl_setopt($ch, CURLOPT_HTTPHEADER, array(
"Authorization: " . $headers['Authorization'],
"Content-type: " . $headers["Content-type"],
"Expect: ")
);
$params = http_build_query($params);
$str = print_r($params, true);
file_put_contents('data_curl1.txt', $str);
curl_setopt($ch, CURLOPT_POST, 1);
curl_setopt($ch, CURLOPT_POSTFIELDS, $params);
$response = curl_exec($ch);
debug($response,"response");
return $response;
}
this is the function with some problems, I try to explain:
I called the oauth_gen passing the parameters array to it, the oauth_gen creates the oauth header that I later used here: "Authorization: " . $headers['Authorization'],.
As I stated, everything is working smoothly, until I have tried to post a gif photoset of 6 files for a total of 6Mb (tumblr permit 2Mb each file and 10Mb total).
PHP runs out of memory and return an error, here it starts my debugging, after a while I contacted the tumblr api helpdesk, and they answer in this way:
You shouldn't need to include the files in the parameters used for
generating the oauth signature. For an example of how this is done,
checkout one of our official API clients.
This changes everything. Untill now, I passed the entire parameters array to the oauth_gen, which, calling the oauth_sig, will rawencode everything into the array (binary strings of gif files inlcuded), with a result of a binary file of about 1Mb becomes at least 3Mb of rawurlencoded string.
and that's why I had memory issues. Nice, so, as the helpdesk say, I've changed the call to the oauth_gen in this way:
$new_array = array();
oauth_gen("POST", "http://api.tumblr.com/v2/blog/$blog/post", $new_array, $headers);
seams legit to me, I passed a new array to the function, the function then generate the oAuth, the headers are passed back and I can use them into the posting call, the result was:
{"meta":{"status":401,"msg":"Unauthorized"},"response":[]}
asking more to tumblr api helpdesk leads only to more links to their documentation and their "tumblr php client" which I can't use, so it isn't a option.
Does anyone has experience with oAuth and can explain me what I'm doing wrong? as far as I understand, the trick is into the encrypted data the oauth_sig create, but I can't figure out how to proceed.
I really want to understand the oauth, but more I read about it and more the tumblr helpdsek seams right to me, but... the solution doesn't work, and works only if I let the oauth function to encrypt the entire data array (with the images and everything) but I can understand that this is wrong... help me.
UPDATE 1
I've tried a new thing today, first I created the empty array, then passed by reference to the oauth_genand only after generating the signature, I've added to the same array all the other fields about the post itself, but the result is the same.
UPDATE 2
reading here: http://oauth.net/core/1.0a/#signing_process
seems that the parameters of the request must all be used for the signature, but this is not totally clear (if someone could explain it better, I really appreciate).
this is weird, because if it's true, it go against the words of the Tumblr help desk, while if it's not true, there is a little confusion in the whole process.
by the way, at this time, I'm stile struck in the same point.
After digging couple of hours into the issue, debugging, reviewing tumblr api and api client, registering a test account and trying to post some images. The good news is finally I come up with a solution. It is not using a native CURL only, you need guzzle and an OAuth library to sign the requests.
Tumblr guys are correct about signing the request. You don't need to pass image data to sign the request. If you check their official library you can see; https://github.com/tumblr/tumblr.php/blob/master/lib/Tumblr/API/RequestHandler.php#L85
I tried to fix the issue with native CURL library but unfortunately I was not successful, either I was signing the request in a wrong way or missing something in the request header, data etc. I don't know actually, Tumblr api is really bad at informing you what you are doing wrong.
So I cheated a little bit and start to read Tumblr api client code, and I come up with a solution.
Here we go, first you need two packages.
$ composer require "eher/oauth:1.0.*"
$ composer require "guzzle/guzzle:>=3.1.0,<4"
And then the PHP code, just define your keys, tokens, secrets etc. Then it should be good to go.
Since the signing request does not include picture data, it is not exceeding memory limit. After signing the request actually we are not getting the contents of the files into our post data array. We are using addPostFiles method of guzzle, which takes care of file addition to POST request, does the dirty work for you. And here is the result for me;
string(70) "{"meta":{"status":201,"msg":"Created"},"response":{"id":143679527674}}"
And here is the url;
http://blog-transparentcoffeebouquet.tumblr.com/
<?php
ini_set('memory_limit', '64M');
define("CONSUMER_KEY", "");
define("CONSUMER_SECRET", "");
define("OAUTH_TOKEN", "");
define("OAUTH_SECRET", "");
function request($options,$blog) {
// Take off the data param, we'll add it back after signing
$files = isset($options['data']) ? $options['data'] : false;
unset($options['data']);
$url = "https://api.tumblr.com/v2/blog/$blog/post";
$client = new \Guzzle\Http\Client(null, array(
'redirect.disable' => true
));
$consumer = new \Eher\OAuth\Consumer(CONSUMER_KEY, CONSUMER_SECRET);
$token = new \Eher\OAuth\Token(OAUTH_TOKEN, OAUTH_SECRET);
$oauth = \Eher\OAuth\Request::from_consumer_and_token(
$consumer,
$token,
"POST",
$url,
$options
);
$oauth->sign_request(new \Eher\OAuth\HmacSha1(), $consumer, $token);
$authHeader = $oauth->to_header();
$pieces = explode(' ', $authHeader, 2);
$authString = $pieces[1];
// POST requests get the params in the body, with the files added
// and as multipart if appropriate
/** #var \Guzzle\Http\Message\RequestInterface $request */
$request = $client->post($url, null, $options);
$request->addHeader('Authorization', $authString);
if ($files) {
if (is_array($files)) {
$collection = array();
foreach ($files as $idx => $f) {
$collection["data[$idx]"] = $f;
}
$request->addPostFiles($collection);
} else {
$request->addPostFiles(array('data' => $files));
}
}
$request->setHeader('User-Agent', 'tumblr.php/0.1.2');
// Guzzle throws errors, but we collapse them and just grab the
// response, since we deal with this at the \Tumblr\Client level
try {
$response = $request->send();
} catch (\Guzzle\Http\Exception\BadResponseException $e) {
$response = $request->getResponse();
}
// Construct the object that the Client expects to see, and return it
$obj = new \stdClass;
$obj->status = $response->getStatusCode();
$obj->body = $response->getBody();
$obj->headers = $response->getHeaders()->toArray();
return $obj;
}
$files = [
"/photo/1.jpg",
"/photo/2.jpg",
"/photo/3.png",
"/photo/4.jpg",
"/photo/1.jpg",
"/photo/2.jpg",
"/photo/3.png",
"/photo/4.jpg",
"/photo/1.jpg",
"/photo/2.jpg",
];
$params = array(
"type" => "photo",
"state" => "published",
"tags"=> [],
"caption"=>"caption",
"link"=>str_replace("_","","http://stackoverflow.com/questions/36747697/oauth-signature-creation-issue-with-php-posting-photoset-to-tumblr"),
"data" => $files,
);
$response = request($params, "blog-transparentcoffeebouquet.tumblr.com");
var_dump($response->body->__toString());
I'm working with an api that returns a number of useful variables within the header response, for example with...
Pragma → no-cache
Server → Apache
X-Pages → 1424
... I'd like to use the X-Pages variable to form some further logic. How do I retrieve this variable and use it within my php script?
My api currently call looks like this:
$username = "XXXXXXXX";
$password = "XXXXXXXX";
$remote_url = 'XXXXXXXX';
$headers = array();
$headers[] = "Authorization: Basic " . base64_encode("$username:$password");
$opts = array(
'http'=>array(
'method'=>"GET",
'header' => $headers
)
);
$context = stream_context_create($opts);
$file1 = file_get_contents($remote_url, false, $context);
Thanks,
Matt
You can find the response headers in $http_response_header after making the call to file_get_contents - it'll be automatically populated in the local scope.
I'm trying to build a basic app where I can modify my own google calendar using Google's RESTful api, but I'm having trouble getting the oAuth token. I've chosen to do it using the Services Application oAuth flow - I don't want to have to constantly re-agree to letting my own app use my calendar, but if there's a better way to do this please let me know.
Every time I make the http request I get a Bad Request error. Any ideas/help?
Here's the code:
<?php
$payload = array(
"iss"=>"services client email",
"scope"=>"https://www.googleapis.com/auth/calendar",
"aud"=>"https://accounts.google.com/o/oauth2/token",
"iat"=>date("U"),
"exp"=>date("U")+3600
);
$key = "Simple API key";
$jwt = encode_header($payload,$key);
print_r(request_g_token($jwt));
function request_g_token($jwt)
{
$data = array(
'grant_type' => 'urn:ietf:params:oauth:grant-type:jwt-bearer',
'assertion' => $jwt
);
$data = http_build_query($data);
$url = "https://accounts.google.com/o/oauth2/token";
//echo $data;
$opts = array('http' =>
array(
'protocol_version' => '1.1',
'method' => 'POST',
'header' => "Host: accounts.google.com\r\n" .
"Content-Type: application/x-www-form-urlencoded\r\n",
'content' => $data
)
);
$context = stream_context_create($opts);
return (file_get_contents($url, false, $context));
}
function urlsafeb64encode($input){
return str_replace('=','',strtr(base64_encode($input),'+/','-_'));
}
function encode_header($payload, $key, $algo = 'RS256'){
$header = array('typ' => 'JWT', 'alg' => $algo);
$segments = array();
$segments[] = urlsafeb64encode(json_encode($header));
$segments[] = urlsafeb64encode(json_encode($payload));
$signing_input = implode('.',$segments);
$sig = sign_encode($signing_input,$key);
$segments[]=urlsafeb64encode($sig);
return implode('.',$segments);
}
function sign_encode($msg, $key){
return hash_hmac('sha256', $msg, $key, true);
}
?>
Any help would be greatly appreciated
UPDATE
So I went through the service process again and realized I need to use a private key, which I'm now doing. My major question is whether or not to include the "private-key.p12" part of what I downloaded from google or not. I'm still receiving a Bad Request error unfortunately...
UPDATE 2
Realized I needed to pull the key from the pk12 file, and I did so with this code:
function getKey($file){
$p12cert = array();
$fd = fopen($file, 'r');
$p12buf = fread($fd,filesize($file));
fclose($fd);
if ( openssl_pkcs12_read($p12buf, $p12cert, 'notasecret') )
{
//worked
$temp = $p12cert['pkey'];
$temp = str_replace("-----BEGIN PRIVATE KEY-----","",$temp);
$temp = str_replace("-----END PRIVATE KEY-----","",$temp);
return $temp;
}
else
{
//failed
return "failed";
}
}
However, it's still giving me a bad request error and I think it's to do with the fact that the key comes back in multiple lines. Any ideas?
For anyone experiencing this issue, I've basically come to the conclusion that google hasn't interfaced calendar with their service accounts yet, and am using a refresh token to achieve similar results without having to log in all the time.