Describe simulator scene with `gpt-4-vision`

[elkraneo]

• Consider using ScreenCaptureKit

• CGWindowListCopyWindowInfo(::) can help to restrict captures to simulator

  // Define a function to capture the screen of the Xcode simulator app window
  func captureSimulatorWindow() {
  let targetAppName = "Simulator"  // The name of the Xcode simulator app
  
  if let windowList = CGWindowListCopyWindowInfo(.optionOnScreenOnly, kCGNullWindowID) as? [[String: AnyObject]] {
      for window in windowList {
          if let windowOwnerName = window[kCGWindowOwnerName] as? String, windowOwnerName.contains(targetAppName) {
              if let windowID = window[kCGWindowNumber] as? CGWindowID {
                  let screenCaptureSession = ScreenCaptureSession(windowID: windowID)
  
                  // Start the screen capture session for the Xcode simulator window
                  screenCaptureSession.startCapture { image, error in
                      if let image = image {
                          // Process the captured image
                          // For example, you can save the image to a file
                          saveImageToFile(image: image)
                      } else if let error = error {
                          // Handle the error
                          print("Error capturing the simulator window: \(error.localizedDescription)")
                      }
                  }
              }
          }
      }
  }
  }

[elkraneo]

Early experimentation here: https://github.com/elkraneo/VisionDescription