/classifier-client-java

Java client for test.ai classifier server

Primary LanguageJava

Test.ai Classifier - Java Client

The code in this directory defines a client library for use with the gRPC-based Test.ai classifier server.

Installation & Setup

At this point, the library is not available on Maven Central for easy download. Instead, check the Releases page to download pre-built Jarfiles you can import into your projects, or use Jitpack.

Usage

This client exposes two classes:

ai.test.classifier_client.ClassifierClient;
ai.test.classifier_client.Classification;

The important class is ClassifierClient, whose constructor takes host and port parameters, so that the client can speak to the correct classifier server. The class has two important methods:

Map<String, Classification> classifyElements(String label, Map<String, byte[]> elementImages,
        double confidenceThreshold, boolean allowWeakerMatches)

List<WebElement> findElementsMatchingLabel (RemoteWebDriver driver, String label,
        double confidenceThreshold, boolean allowWeakerMatches)
  1. classifyElements takes a label (see lib/labels.js in this repo), a map of Strings (ids) to byte arrays (representing PNG image data), a confidence threshold (1.0 = perfect confidence required for a match, 0.0 = no confidence required), and a boolean flag which tells the server whether or not to return potential matches even if the potential match had a different label as its highest-confidence classification. The return value is a map of Strings (the same ids you passed in) to Classification objects (described below).
  2. findElementsMatchingLabel is a helper function for use with Selenium tests (for Appium use the Appium plugin as described in the main README for this repo). It takes a driver object and the same final parameters as classifyElements. The plugin will use the driver object to take screenshots of relevant images and pass them to the classifier. The return value is a list of any WebElements which match the label provided.

The Classification object is simply a container for the label, confidence, and confidence for the label which was originally provided.

For a concrete example, check out the ClientSeleniumTest.java file in this repo.