An Interest In:
Web News this Week
- April 25, 2024
- April 24, 2024
- April 23, 2024
- April 22, 2024
- April 21, 2024
- April 20, 2024
- April 19, 2024
Perspective API
What is Perspective?
Perspective is a free API that uses machine learning to identify "toxic" contents, making it easier to organize better online conversations.
By scoring a sentence based on the perceived impact the text may have in a conversation, developers and editors can use this score to provide feedback to commenters, help moderators review comments more easily, or help readers filter out "toxic" language. Perspective provides scores for several attributes, such as:
- Severe toxicity
- Insults
- Injuries
- Identity attacks
- The threats
- And the sexually explicit
So we have a very good solution to protect ourselves from threatening comments and so on to avoid broadcasting bad content whether it is in a comment, in a forum or in a live chat. There are multitudes of functionality to realize with this API.
Perspective API is a very good way to filter out insults and phrases that are toxic.
Toxicity online poses a serious challenge for platforms and publishers. Online abuse and harassment silence important voices in conversation, forcing already marginalized people offline.
An example of code
The example is under JavaScript, but can totally work with any other language.
// Since the official documentation and slightly modified // https://developers.perspectiveapi.com/s/docs-sample-requestsconst {google} = require('googleapis')require('dotenv').config()CONTENT = 'You're really crap at this game'// Create an .env file to recover GOOGLE_API_KEY.API_KEY = process.env.GOOGLE_API_KEYDISCOVERY_URL = 'https://commentanalyzer.googleapis.com/$discovery/rest?version=v1alpha1'google.discoverAPI(DISCOVERY_URL).then(client => { const analyzeRequest = { comment: { text: CONTENT, }, requestedAttributes: { TOXICITY: {}, }, } console.info(`Input Text : ${analyzeRequest.comment.text}`) client.comments.analyze( { key: API_KEY, resource: analyzeRequest, }, async (err, response) => { if (err) throw err let ScoreValue = response.data.attributeScores.TOXICITY.summaryScore.value await console.log(`TOXICITY Score : ${ScoreValue}`) console.table(JSON.stringify(response.data, null, 2)) })}).catch(err => { throw err})
It's so easy to set up, just install googleapis
and dotenv
and get your Perspective API key and test the code.
yarn add googleapis dotenv
Or if you prefer NPM :
npm i googleapis dotenv
I have already made a project that is Open Source on GitHub called No Toxic Discussions. It's an GitHub Action that identifies the message in the discussion space and checks if its content is toxic or not.
thomasbnt / actions-NoToxicDiscussions
No Toxic Discussions, a GitHub Action to detect toxicity in discussions area.
You have also a DEV post about this Action. Click here to read him.
Credits
Some texts have been copied from the website, as for the banner the logo "Perspective API" comes from Jigsaw of Google Inc. The source code has been taken from the example on their website and modified so that the result is visible. It comes from the modification I made for the No Toxic Discussions project.
Check my Twitter account. You can see many projects and updates. You can also . |
---|
Original Link: https://dev.to/thomasbnt/perspective-api-20al
Dev To
An online community for sharing and discovering great ideas, having debates, and making friendsMore About this Source Visit Dev To