´´I later realise the requirement of proper commit history. Considering that I have been working on this task locally I have tried to divide it into commits but not usually how I go about commiting. You will see all the commits on one date cos of that reason.´´
Create a service for storing and retrieving image data. The service should follow best practices of a web service, examples:
- data validation
- correct status codes
- proper headers (like
Content-Type, etc) - ...
Feel free to use whatever database backend you like. For programming language we would like Go to be used.
The service should be performant and ideally easily scalable. To limit the scope of the work it's OK to make less than ideal choices as long as they are motivated. For instance "I used the Windows Registry on my computer as database, it won't scale beyond three entries but if I wanted to make it scalable I would change to <insert buzzword here>".
We want the code with a nice commit history as a git bundle (git bundle create <bundle.name> --all) and if you can deploy it somewhere that would be great but it's not a requirement. Any questions can be sent to gustaf@apple.com.
GET /v1/images
List metadata for stored images.
GET /v1/images/<id>
Get metadata for image with id <id>.
GET /v1/images/<id>/data
Get image data for image with id <id>.
Optional GET parameter: ?bbox=<x>,<y>,<w>,<h> to get a cutout of the image.
POST /v1/images
Upload new image. Request body should be image data.
PUT /v1/images/<id>
Update image. Request body should be image data.
Image metadata should be the following fields:
- Filesize of the image
- Image dimensions
- Image type (gif, jpg etc)
- Date/time image was uploaded
- Extra fields of your choice
It is up to you how you want to represent the image metadata in your API.
Some notes about the assignment:
To test update: curl -X PUT -F 'file=@/path/to/image/file' localhost:8089/v1/images/:id -H "Content-Type: multipart/form-data"
In ideal scenarios we should convert all the images into one of png/ jpg to make code cleaner and handling of images much easier. Many tools like ffmpeg can be a good choice to convert every image to certain types depending on the requirement of course.