A Javascript client for the Perspective API
A simple example JS/TS client library for the Perspective API.
This client can be used to send text to the API and get back the scores for a set of model attributes.
This libray is published on npm as @conversationai/perspectiveapi-js-client.
To use it, you'll first need an API key from ConversationAI's Perspective API. Then you can use the library like this:
``
const { Client } = require("@conversationai/perspectiveapi-js-client");
const client = new Client(YOUR_API_KEY);
const results = client.getScores(text, options);
This client exposes one method, getScores.
getScores(text, options);
text (Required)
A string between 1 and 3000 characters representing a piece of text (e.g. a comment on an article) to be analyzed by Perspective API.
options
An optional object that specifies options for the API.
Accepted keys in the options object are:
Field | Description
----- | -----------
context.entries | A list of objects providing the context for text. Defaults to null, equivalent to an empty list.context.entries[].text | The text of a context object.attributes | An array of strings specifying the attributes to get scores for. See the Perspective docs for a list of available attributes. Default: ["SPAM", "TOXICITY"] | A list of ISO 631-1 two-letter language codes specifying the language(s) that text is in (for example, "en", "es", "fr", etc). Default: ["EN"] | Whether the API is permitted to store text and context from this request. Default: true | A boolean specifying whether to strip html tags from text. Default: true
A promise that rejects with an Error object containing a message property specifying the error, or resolves to an object of attribute scores of the shape [attribute: string]: number, where the number is a float between 0 and 1. Here's an example result:`
{
TOXICITY: 0.65,
SPAM: 0.25
}
The above result would mean that the Perspective API suggests the text is 65% likely to be toxic and 25% likely to be spam.
Run yarn test` to run the test suite.