OGP service

Usage instructions

Note: There is a postman collection included in the tar file if you prefer using postman.

Here are some example requests:

POST request example # returns { 'id': "uuid" } or error

curl -X POST \
  'http://marius.pop.hiring.keywee.io/stories?url=https://ogp.me/' \
  -H 'cache-control: no-cache'

GET request example # returns parsed OGP tags or error

Note: Replace id with id received from the request above

curl -X GET \
  http://marius.pop.hiring.keywee.io/0e537896-d1ad-4add-83b3-633389624b32 \
  -H 'cache-control: no-cache'

Functionality

The Object Type tags mentioned here and below, have not been scraped. Only "og:" prefixed tags are scraped at this time but the pattern can be easily extended for other tags.

Development Setup instructions

install and start redis
install ruby 2.7.2
run: bundle install
run: bundle exec puma -C config/puma.rb

Start puma server on provided server

RACK_ENV=production bundle exec puma -C config/puma.rb --bind unix:///tmp/puma.sock

Run tests

bundle exec rspec --format documentation

Other thoughts and considerations

Redis was used for the purpose of this exercise as an in memory data store and to satisfy the single scraping constraint. The assumption was that only one Redis instance would be running that would handle race conditions from the puma servers. A persistent NoSQL data store would likely be preferable for the long term storage.

There's an argument to be made against the additional complexity added by handling race conditions to ensure the single scrape requirement is satisfied.