Add WEB_DETECTION to Nuxeo Vision (supported by Google)

Since recently, Google Vision added the WEB_DETECTION tag for images. it returns info on "same image", "partially same" but also list of pages using the image etc.
See: https://cloud.google.com/vision/docs/reference/rest/v1/images/annotate#WebDetection

Unfortunately, developers can't use the new tag because these tags are enums in the code, so trying to use a non existant label throws an IllegalArgumentException when calling VisionOp operation (https://github.com/nuxeo/nuxeo-vision/blob/df2b2e298ab99f2c8d9dcf2dc073ddceef452146/nuxeo-vision-core/src/main/java/org/nuxeo/vision/core/operation/VisionOp.java#L88)

So, first, let's add "WEB_DETECTION" to VisionFeatures

Reminder: Nuxeo Vision does not encapsulate all and every feature in sub-methods. Caller has to use the getNativeObject() and handle the result from this JSON object. In JS Automaiton, something like:

. . .
blob = VisionOp(blob, {
  features: ['WEB_DETECTION'],
  maxResults: 20,
  outputVariable: 'results'
});
nativeResult = ctx.results.get(0).getNativeObject();  

nativeResult.webDetection.webEntities.forEach(function(entity) {
 Console.log("" + entity.score + " • " + entity.description );
});

nativeResult.webDetection.fullMatchingImages .forEach(function(entity) {
  Console.log("" + entity.score + " • " + entity.url );
});

. . . etc . . .

Also notice AWS Rekognition does not support this feature (but it's OK, it does not support all the same featires as Google Vision already)