Package Exports

perspective-api-client

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (perspective-api-client) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

perspective-api-client

NodeJS client library for the Perspective API.

Install

$ npm install perspective-api-client

Usage

const Perspective = require('perspective-api-client');
const perspective = new Perspective({apiKey: process.env.PERSPECTIVE_API_KEY});

(async () => {
  const text = 'What kind of idiot name is foo? Sorry, I like your name.';
  const result = await perspective.analyze(text, {doNotStore: true});
  console.log(JSON.stringify(result, null, 2));
})();
// {
//   "attributeScores": {
//     "TOXICITY": {
//       "spanScores": [
//         {
//           "begin": 0,
//           "end": 56,
//           "score": {
//             "value": 0.8728314,
//             "type": "PROBABILITY"
//           }
//         }
//       ],
//       "summaryScore": {
//         "value": 0.8728314,
//         "type": "PROBABILITY"
//       }
//     }
//   },
//   "languages": [
//     "en"
//   ]
// }

Specifying models

The TOXICITY model is used by default. To specify additional models, pass options.attributes.

(async () => {
  const text = 'What kind of idiot name is foo? Sorry, I like your name.';
  const result = await perspective.analyze(text, {attributes: ['toxicity', 'unsubstantial'], doNotStore: true});
  console.log(JSON.stringify(result, null, 2));
})();
// {
//   "attributeScores": {
//     "TOXICITY": {
//         ...
//       }
//     },
//     "UNSUBSTANTIAL": {
//       "spanScores": [
//         {
//           "begin": 0,
//           "end": 32,
//           "score": {
//             "value": 0.72937065,
//             "type": "PROBABILITY"
//           }
//         },
//         {
//           "begin": 32,
//           "end": 56,
//           "score": {
//             "value": 0.8579436,
//             "type": "PROBABILITY"
//           }
//         }
//       ],
//       "summaryScore": {
//         "value": 0.6332942,
//         "type": "PROBABILITY"
//       }
//     }
//   },
//   "languages": [
//     "en"
//   ]
// }

More options

You can also pass an AnalyzeComment object for more control over the request.

(async () => {
  const text = 'What kind of idiot name is foo? Sorry, I like your name.';
  const result = await perspective.analyze({
    comment: {text},
    requestedAttributes: {TOXICITY: {scoreThreshold: 0.7}}
  });
  console.log(JSON.stringify(result, null, 2));
})();
// {
//   "attributeScores": {
//     "TOXICITY": {
//       "spanScores": [
//         {
//           "begin": 0,
//           "end": 56,
//           "score": {
//             "value": 0.8728314,
//             "type": "PROBABILITY"
//           }
//         }
//       ],
//       "summaryScore": {
//         "value": 0.8728314,
//         "type": "PROBABILITY"
//       }
//     }
//   },
//   "languages": [
//     "en"
//   ]
// }

API

perspective = new Perspective()

analyze(text, [options])

text

Type: String or Object

Either the text to analyze or an AnalyzeComment object.

options

attributes

Type: Array or Object

Model names to analyze. TOXICITY is analyzed by default. If passing an Array of names, the names may be lowercased. See https://github.com/conversationai/perspectiveapi/blob/master/api_reference.md#models for a list of valid models.

doNotStore

Type: Boolean Default: false

Whether the API is permitted to store comment and context from this request.