ml-news-article-scraper-example
In this example, we will see how we can work with Machine Learning applications written in Python with a NodeJS Script, to build a Polyglot Machine Learning application for scraping similar news articles.
Install
Install MetaCall CLI:
$ curl -sL https://raw.githubusercontent.com/metacall/install/master/install.sh | sh
Install application dependencies:
- For Python:
metacall pip3 install -r requirements.txt - For NodeJS:
metacall npm i readline-sync
Run the Example
$ metacall app.js
Once the application is kick-started, you will be prompted to enter a News Article which you would like to find similar articles for. Let's use this sample article for testing our application: https://www.nytimes.com/2021/03/23/business/teslas-autopilot-safety-investigations.html
Here is the application output:
$ metacall app.js
Information: Global configuration loaded from /gnu/store/5cxmq6y8z24ijnvhh6lndgpriwnhf3jl-metacall-0.3.17/configurations/global.json
Enter the News URL:
https://www.nytimes.com/2021/03/23/business/teslas-autopilot-safety-investigations.html
┌─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┬───────────────┐
│ (index) │ Values │
├─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┼───────────────┤
│ https://auto.timesofindia.com/news/others/teslas-autopilot-technology-faces-fresh-scrutiny/articleshow/81652823.cms │ '83.68405286' │
│ https://www.autosafety.org/teslas-autopilot-technology-faces-fresh-scrutiny/ │ '60.35694007' │
│ https://www.anandmarket.in/teslas-autopilot-technology-faces-fresh-scrutiny/ │ '94.97681053' │
│ https://www.entrepreneur.com/article/367724 │ '60.67538891' │
│ http://www.newsnetworks.in/india/teslas-autopilot-technology-faces-fresh-scrutiny/ │ '0.' │
└─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┴───────────────┘
Script (app.js) loaded correctly
Deployment using MetaCall FaaS
After deploying the application into the FaaS https://dashboard.metacall.io, it can be accessed with (change <your_alias> by the alias you used to sign up):
curl -X POST https://api.metacall.io/<your_alias>/ml-news-article-scraper-example/v1/call/similarNews -X POST --data '{ "url": "https://www.nytimes.com/2021/03/23/business/teslas-autopilot-safety-investigations.html" }'
