php-crowdflower

A toolkit for interacting with CrowdFlower via the REST API.

I've used this in production for a period of time and had it handling jobs costing $100s of dollars. Although I'm not using it in production at the moment, the tests (which not only test this PHP API but the CrowdFlower REST API by extension) have running in Hudson (code checked directly out of github) daily since 4/16/2011. Currently they are passing, with the stats being: Tests: 76, Assertions: 86, Incomplete: 3.

Example Usage

Fetch a particular job:

    $job = CrowdFlower::factory("job", $apiKey);
    $resp = $job->get($jobID);

Upload a CSV file for a job:

    $upload_data['file'] = '/path/to/my/file'
    $upload_data['content-type'] = "text/csv";
    //if appending, you need the id, otherwise omit it
    $upload_data['id'] = "text/csv";
    $job = CrowdFlower::factory("job", $apiKey);
    $resp = $job->upload($upload_data);

Copy an existing job into a new one:

    $job = CrowdFlower::factory("job", $apiKey);
    $copy_resp = $job->copy($jobID, array('all_units' => 'true'));
    //get new job id
    $copied_job_id = $copy_resp['response']->id;
    print $copied_job_id;

Check the status of a job:

    $job = CrowdFlower::factory("job", $apiKey);
    $job->ping($jobID);

Order Units in a job

    $job = CrowdFlower::factory("job", $apiKey);
    //$channels one or more of: array('amt', 'iphone', 'mob', 'sama')
    //mob is free and is for testing, others require money in your account
    $channels = array('mob', 'amt');
    $job->order($jobID, $unitsToOrder, $channels);

Download all judgments from a job

    $job = CrowdFlower::factory("job", $apiKey);
    $filename = '/path/to/download';
    //the file will be downloaded to the path above
    //$resp will have the API resp
    $resp = $job->download($jobID, $filename, true);

General Structure of the Response Object

When making calls to the API, an array is returned with the following:

1. 'response', which contains the actual json_decoded response from Crowdflower
2. 'info', which contains the associative array returned by curl_getinfo()
3. If an error due to curl (network timeout, etc.) occurs, the array will contain two extra elements: error_msg and error_code
4. 'request', the complete request, including URL, method, and data sent

Contributing

1. Fork php-crowdflower
2. Create a topic branch - git checkout -b my_branch
3. Make your feature addition or bug fix and add tests for it.
4. Commit, but do not mess with the rakefile, version, or history.
5. Push to your branch - git push origin my_branch
6. Create an Issue with a link to your branch

Copyright

Copyright © 2011 Tom Melendez. See LICENSE for details.