I am in the process of developing a web-based RESTful api for a client. Everything works great apart from one request, in which I need to request the Foursquare API for each row.
The URL for this request is: http://api.example.com/v1/users/times.
Currently the response of a request to that url is:
{
"response": {
"user": {
... some user info ...
"times": [
{
"id": "8",
"venue_fq_id": "4b81eb25f964a52000c430e3",
"user_id": "1",
"wait_length": "4468",
"created_at": "2012-06-09 21:45:43"
},
{
"id": "9",
"venue_fq_id": "4aad285af964a520c05e20e3",
"user_id": "1",
"wait_length": "8512",
"created_at": "2012-06-09 21:45:43"
},
{
"id": "10",
"venue_fq_id": "42377700f964a52024201fe3",
"user_id": "1",
"wait_length": "29155",
"created_at": "2012-06-09 21:45:44"
},
{
"id": "11",
"venue_fq_id": "45c88764f964a5206e421fe3",
"user_id": "1",
"wait_length": "33841",
"created_at": "2012-06-09 21:45:44"
},
{
"id": "12",
"venue_fq_id": "430d0a00f964a5203e271fe3",
"user_id": "1",
"wait_length": "81739",
"created_at": "2012-06-09 21:45:44"
}
]
}
},
"stat": "ok"
}
However, the venue_fq_id being returned in the response.user.times array is relative to a venue on the Foursquare API. I tried running a curl request to the Foursquare API for each row, but the performance is incredibly slow. Please can you give some examples of ways in which I could speed up the performance while retrieving the same information I would had I accessed requested the F/Q API each time?
Here’s my code:
$query = $this->db->query("SELECT * FROM `wait_times` WHERE `user_id` = ?", array($email_address));
$wait_times = $query->result();
foreach ($wait_times as $wait_time) {
$wait_time->venue = $this->venue_info($wait_time->venue_fq_id);
}
function venue_info($fq_id) {
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, "https://api.foursquare.com/v2/venues/4b522afaf964a5200b6d27e3");
curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
$response = json_decode(curl_exec($ch));
curl_close($ch);
return $response['response']['venue'];
}
You’re wasting a horrendous amount of time instantiating/tearing down CURL objects. This prevents you from taking advantage of HTTP keep-alives, forcing curl to start a new tcp connection for every request you make.
Curl handles CAN be reused. e.g.