Cloud Batches


  • This tutorial assumes that RAPP API is installed and built
  • This tutorial uses the 0.7.0 version of the C++ API

In the following example we are going to execute three actions at the same time:

  • face detection
  • human detection
  • door angle detection

Include the following headers:

  • service_controller.hpp
  • picture.hpp
  • vision_detection.hpp
  • iostream

Then create a service controller as in the previous examples:

rapp::cloud::platform info = {"rapp.ee.auth.gr", "9001", "rapp_token"}; 
rapp::cloud::service_controller ctrl(info);

In this example we are going to use vision detection, so we a picture. We will try to detect faces, humans and doors in the same picture. If loading a picture from disk, then constuct a rapp::object::picture object (replace my_picture.jpg with a valid one!).

NOTE: only PNG and JPEG are formats supported currently.

auto pic = rapp::object::picture("my_picture.jpg");

Each cloud call takes a different callback functor, so we'll create one for each call. You can find more information for each class in rapp/cloud/vision_detection.hpp

  • face detection:
auto face_cb = [&](std::vector<rapp::object::face> faces) { 
    std::cout << "Found " << faces.size() << " faces!" << std::endl;
};
  • human detection:
auto human_cb = [&](std::vector<rapp::object::human> humans) {
    std::cout << "Found " << humans.size() << " humans!" << std::endl;
};
  • door angle:
auto hazard_cb = [&](double door_angle) {
    std::cout << "Door angle: " << door_angle  << std::endl;
};

When we run the batch, we construct the cloud call in inline object as we execute service_controller::make_calls:

ctrl.make_calls(rapp::cloud::face_detection(pic, false, face_cb),
                rapp::cloud::human_detection(pic, human_cb),
                rapp::cloud::door_angle_detection(pic, hazard_cb));

When the platform responds, we'll receive:

Door angle: 2
Found 2 humans!
Found 1 faces!

The order of the calls may change: there is a possibility that face detection may finish first and then human detection.