Test.ai Classifier - Node.js + WebdriverIO Client

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

Installation & Setup

npm install test-ai-classifier-client

Usage

This client exposes as the default export a ClassifierClient class, which can be instantiated as follows:

const ClassifierClient = require('test-ai-classifier-client');
const client = new ClassifierClient({host, port})

(Where host and port refer to the address and port the server is running on).

The client exposes two instance methods:

1. classifyElements takes an object parameter with 4 keys:

   1. labelHint: the label you wish to find matching elements with (see lib/labels.js in this repo).
   2. elementImages: an object whose keys are ids, and whose values are Buffer objects containing raw binary data of PNG images.
   3. confidenceThreshold: (optional) the confidence below which not to return matches (0.0 - 1.0)
   4. allowWeakerMatches: (optional) whether or not to return a match for elements that did match the label, but for whom another label had a higher confidence.

   The return value of this method is an object whose keys are the same ids you sent in, and whose values are classification objects with the following keys:

   1. label: the matching label
   2. confidence: the confidence for the matched label
   3. confidenceForLabel: the confidence for the label given in labelHint

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 an object parameter with 4 keys: all of the same keys as in classifyElements except for elementImages which is replaced by:

   1. driver: the WebdriverIO driver object The return value of this method is an array of WebdriverIO element objects that match.

For a concrete example, see test/rpc-e2e-specs.js

Development

Tests expect a running Classifier server on port 50051.