This tutorial assumes that you have already created and setup your Cognitive Services as well as Easy Tables in your Azure Portal If you have not, please start by doing so.
To be able to make REST requests to Azure/Cognitive Services, we will be using a part of Unity's built in WWW library. More info can be found at https://docs.unity3d.com/ScriptReference/WWW.html.
To start off, we need to recognize the request format that cognitive services accepts. In the case for the Computer Vision API, we need to include Ocp-Apim-Subscription-Key, Content-Type as well as the correct url that we're hitting. In this case we use https://api.projectoxford.ai/vision/v1.0/analyze?visualFeatures=Categories with parameters already added to the url.
To get your Ocp-Apim-Subscription-Key, go to your Azure Portal and select your cognitive services Api that you have already created and navigate to keys under Resource Management. You will find two keys; you will be able to use either.
//Make a new form
WWWForm form = new WWWForm();
//Create and attach headers
var headers = form.headers;
headers["Ocp-Apim-Subscription-Key"] = "999999999xxxxxxxxxyourkeygoeshere";
headers ["Content-Type"] = "application/octet-stream";
Your Content-Type will be different depending on the service. In this case, we analyzed an image that was captured, so we used application/octet-stream.
Note: Do NOT use form.AddBinaryData() and form.AddField() to add the body of the request, we have found them to overwrite the content type in some of the cases. To overcome this, we will be feeding binary data straight into the www request function.
//Declare byte array of image to be sent
byte[] imageData = null;
//You can either add straight into function or pull from storage
FileInfo fileInfo = new FileInfo("pic.jpg");
long imageFileLength = fileInfo.Length;
FileStream fs = new FileStream("pic.jpg", FileMode.Open, FileAccess.Read);
BinaryReader br = new BinaryReader(fs);
//Map image into byte array
imageData = br.ReadBytes((int)imageFileLength);
You can now declare your url as a string and make the request.
string url = "https://api.projectoxford.ai/vision/v1.0/analyze?visualFeatures=Categories";
WWW www = new WWW(url, imageData, headers);
Use IEnum function to get response. StartCoroutine(WaitForRequest(www));.
IEnumerator WaitForRequest(WWW www)
{
yield return www;
// check for errors
if (www.error == null)
{
//Do something with response
Debug.Log("WWW Ok!: " + www.text);
} else {
Debug.Log("WWW Error: "+ www.error);
}
}
More info can be found at https://dev.projectoxford.ai/docs/services/56f91f2d778daf23d8ec6739/operations/56f91f2e778daf14a499e1fa.
The following headers have to be included in the request:
headers["ZUMO-API-VERSION"] = "2.0.0" ;
headers["Content-Type"]="application/json";
Add the Body in json format. Multiple items can be added.
string newObj = "{\"newTitle\":\"My Title\"}";
If object(column) is not present in table, it will be added.
define url
string url = "http://account_name.azurewebsites.net/tables/table_name";
since www request only takes in byte array as a body, the json will have to be cast into a byte array through Encoding.ASCII.GetBytes(newObj) this uses System.Text.
looking at it in the request, we get WWW www = new WWW(url, Encoding.ASCII.GetBytes(newObj), headers);.
Use IEnum function to get response. StartCoroutine(WaitForRequest(www));.
IEnumerator WaitForRequest(WWW www)
{
yield return www;
// check for errors
if (www.error == null)
{
//Do something with response
Debug.Log("WWW Ok!: " + www.text);
} else {
Debug.Log("WWW Error: "+ www.error);
}
}
To find out more about how to insert into a table, check out: https://msdn.microsoft.com/en-us/library/azure/dd179433.aspx
To find out more about how to query, check out: https://msdn.microsoft.com/en-us/library/azure/dd179421.aspx
Note: Make sure your account does not have app service authentication on (in Azure portal) or else you will have to add a header containing your key.
##If you would like to add onto this tutorial, do a pull request!