xywzel/fileserver

Recruitment task of file sharing REST API using Django

FileServer

Recruitment task of file sharing REST API using Django

The assignment was as follows:

''' Please write a web application that provides a REST API for logged-in users to upload and download any kind of files.

The users must be able to login and logout (use either token or session authentication). There is no need to implement CRUD endpoints for users or organizations, those can be created by running a script.

Each user and file must belong to an organization. Once uploaded, the file must belong to the same organization as the user who uploaded it.

Users should see and be able to download any of the uploaded files from any organization. Write an endpoint for listing all the files that belong to one organization, and an endpoint for listing all the file downloads done by one user. Include timestamps when the file was uploaded and when the user downloaded a file.

Keep track of how many times each file has been downloaded, and also how many total file downloads each organization has (number of all file downloads from that organization). Include the number of downloads when listing files and organizations.

Use Django and Django REST Framework (https://www.django-rest-framework.org/). You can use a database of your choice.

No need to implement the UI, writing the API endpoints is enough. '''

Scripts

• setup.sh makes initial setup of the environment
• clean.sh removes existing databases and migrations, allowing starting testing from clean slate. It also runs create.py that initializes some organizations and users with their profiles
• start.sh runs the development server in its virtual environment

Implementation

• /files/ is root for the API
• Users are handled with django auth User model, extended with Profile model. Profiles can be queried through /files/profile/.
• Organizations can be queried trough /files/organization/. Organizations track their total download count and have reverse references to member profiles, uploads and records.
• Files can be queried trough files/upload/, the serialised format include a download link for the file in data field. Name is upload, because file is too overloaded word in python and django context. Uploads track their total download count.
• DownloadRecords allow for more granular tracking of downloads, each recording, user, organization, file upload and time of download. With filtering these, each of the required views should be possible to display. They can be accessed trough /record/
• The rest API uses hyperlinked references, so other models are referenced by a link that fetches information about that model.
• Currently there is no proper way to upload files, way to do it trough form was left unfinished and I did not have time to find a way to do it trough the serialised API, though it should be possible. Files can be added trough admin interface admin/files/upload/add/.
• There is no consideration of security or privacy currently in the system as all information that is visible is so for all logged in users and can likely also be edited trough POSTs to the API.

About the workload

• This took about two work days (~16h) split around 3 days.
• Most of this time was spend reading django and REST framework documentation and tutorials.
• Trying to get the database play nicely also took quite a while, and is the reason, why my clean up script takes quite drastic measures.
• Major improvement left to do would have been ability to filter DownloadRecords in queries based on profile, file or organization. This would have taken additional views or changes to make current ones take parameters. Though including reverse records for user and organization queries seems to do the same.