https://github.com/testdotai/classifier-client-node

Node.js client for test.ai classifier server
https://github.com/testdotai/classifier-client-node

Last synced: about 2 months ago
JSON representation

Node.js client for test.ai classifier server

• Host: GitHub
• URL: https://github.com/testdotai/classifier-client-node
• Owner: testdotai
• Created: 2019-12-19T00:53:39.000Z (about 5 years ago)
• Default Branch: master
• Last Pushed: 2022-04-09T04:41:04.000Z (over 2 years ago)
• Last Synced: 2024-09-15T22:40:47.638Z (4 months ago)
• Language: JavaScript
• Size: 1.48 MB
• Stars: 7
• Watchers: 4
• Forks: 3
• Open Issues: 12
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# 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](https://github.com/testdotai/appium-classifier-plugin).

## 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:

```js

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.