This API application uploads documents to a folder and follows the specification outlined.
Navigate to the content of the go
git clone https://github.com/osazemeu/shiny-octo-meme.git
cd shiny-octo-meme
Ensure you have Postgres instance running locally in server and run the following command
rails db:setup
rails test
rails serve
navigate to the following link on your browser http://localhost:3000/ # Yay! You're on Rails!
Screenshot: https://imgur.com/baNx4b1
curl -X POST \
http://localhost:3000/api/files \
-H 'Content-Type: multipart/form-data' \
-H 'Host: localhost:3000' \
-F name=osazeme \
-F 'tags[]=zync' \
-F 'file=@/Users/osazemeusen/Downloads/Amazon Web Services Ackenowledgement.pdf' \
-F 'tags[]=golem' \
-F 'tags[]=Phoenix' \
-F 'tags[]=syndicate' \
-F 'tags[]=drawl' \
-F 'tags[]=stinger'Screenshot: https://imgur.com/IMmgCsj
curl -X GET \
'http://localhost:3000/api/files/+golem%20+zync%20-stinger/0' \
-H 'Host: localhost:3000'I went with Service Oriented Architecture where services are used for enhanced separation of concern. For uuid, I used PostgreSQL uuid for to generate id. To enhance performance for queries relating to the tag field, I used GIN index. For file uploads, I used ActiveStorage and implemented rollback and error message for validation of files. I intentionally didn't include activerecord-serializer because requirements were pretty straightforward, but it could be included on request. Error handling was done by extending existing functionality in rails.
Write more unit test and integration test, etc to improve coverage. As the application scales with more users and similar search queries, we could use redis to aggressively cache searches for faster retrievals. I didn't include it because I felt the cache would be rewritten rather often (i.e. more writes than reads would invalidate most queries), causing thrashing.
8.32hrs