Setup the server

1. Sync the DB and create a superuser
2. Populate the DB with some CSV files. There are two ways to do it:

• From the admin panel
• Make a get request to http://127.0.0.1:8000/populate (only available for testing purposes). It adds two example CSV files. Avoid using chrome to access the URL since it makes one query when url is type and another when it's submitted so it may create duplicates

Note: If the CSV files are added from the admin panel, the first time the images are requested it will take more time since all of them need to be cached by first time.

How to use the API

To use the API simply make a GET request to the following URL:

• Cached access http://127.0.0.1:8000/images

It is possible to force the server to check if there are changes in the sources and bring their content to the server.

• Force server to check updates http://127.0.0.1:8000/images?f

Only the entries that can be succesfully load and pass the validation will be returned by the API to the clients. A log is generated with the image errors to be fixed by the suppliers (see the Logs secction below)

Output formats

Beside the required JSON, the API also can be served with a HTML browsable interface. To select between formats use the sufixs .api and .json

Adding more CSV files

There is no limit for the number of source files. The can be added and removed from the admin panel (for that is needed to create a superuser)

Examples:

• http://127.0.0.1:8000/images.api
• http://127.0.0.1:8000/images.json?f

Caching images

When the images are downloaded by first time they will be optimized and stored in the server.

Everytime a SCV file is fetched, the file is hashed and the hash is stored. It will be used to check whether the CSV file changed since the last time it was downloaded.

Assumptions

• If a CSV file is completely removed, the server will keep the entries in the DB for backup purposes. The entries can be removed from the database:
• Removing all the entries from the CSV file (but leaving the file with header)
• Removing the CSVFile entry from the database

Compressing images

For the sake of example, the following are the weights of the default set of images with and without compresion

• Original set of images 10.8Mb
• Compressed set of images 2.33Mb (Saving 78%)

Logs

A log file is generated with the source errors to be submitted to the image supplier to fix them. The file is located in logs/image_suppliers_log.txt

...
[11/09/2015 10:38:39] INFO The image didn't pass the validation: ['The image has no title, which is required. (URL: https://avatars2.githubusercontent.com/u/11301498?v=3&s=200)']
[11/09/2015 10:38:39] INFO The image didn't pass the validation: ['The image has no title, which is required. (URL: https://avatars2.githubusercontent.com/u/11301498?v=3&s=200)']
[11/09/2015 10:38:39] INFO The image didn't pass the validation: ['The URL is not correctly formatted. Enter a valid URL (URL: asdsa)']
...

Production server

I was setting up the production server with Heroku when I realized that it does not support storeing files in it's server. Currently the app generated stores the images in the server so it's not working properly.

The server is running, but it does not cache the images:

• https://quiet-wave-8914.herokuapp.com/images.json (The first time run it's a bit slower since it's a free account in Heroku and the server is sleeping when not used)

The solutions could be:

• Serving it from a VPS where you are free to store your content. But this option would requite more effort.
• Simply using a S3 (i.e. Amazon S3) to store the images.

I'm not sure if setting the production server is something that is expected for me to do but if so, please let me know and I will gladly do it.