The intention of this page is to give a comprehensive and up to date overview over our current system (a task that is bound to fail… Everybody feel free to contribute).
Rough overview
Fashion Recommendations
<HTML><ol></HTML>
Customer sends us a feed with product information (mandatory fields are:
<HTML><ol></HTML>
a unique id that is used to query recommendations,
a category to segment the product catalog,
a link to an image, that depicts the product)<HTML></ol></HTML>
We parse the feed to extract optional meta data and process the relevant images (feed import and enrichment)
The data from the previous step has to be stored and pushed to the frontends
The frontends process queries from our customers and return recommendations
each request is sent to a billing server, where it is further processed for statistics and “long term” storage<HTML></ol></HTML>
Visualytics (Backend for Fashion Recommendations)
<HTML><ol></HTML>
A web based GUI (written in PHP) allows our customers to view some details and make some modifications to some settings
the GUI communicates with a web service (written in python+flask) to retrieve information and to make changes to the system