At the moment the html is accessed from the cache, then scraped. This leads to long loading times, even though a cache is used. Instead the data itself should be cached and html should only be requested and scraped if the cache exceeds the TTL.

After a quick look at the current stable v2 API, I found empty food category-objects.

{
    "type": "",
    "meal": [
        
    ]
},
{
    "type": "",
    "meal": [
        
    ]
},
{
    "type": "",
    "meal": [
        
    ]
}

full day's result @ pastebin
full day's result @ hastebin

There are still mutiple spaces instead of a single space in the dish name. There is a regex that should filter out double spaces, but it does not seem to work for all cases.

There are different drinks provided in the Mensa, these could be statically served aswell from e.g. a json file that is read.
Access to the Mensa Bremen is needed to see these prices.

The Studentenwerk website https://www.stw-bremen.de/de/essen-trinken/uni-mensa shows Symbols according to certain ingredients like Pork, Fish or a Vegan Mention. These could be implemented into the API aswell, providing more information about the different meals.

Aufläufe & Gratin, Beilage, Suppen and Salate have multiple options. They have to be seperate in the response.

On accessing the API the program throws 2 unhandled promise rejection errors. These should be handled.

For now the API only serves costs for students. The Studentenwerk Website https://www.stw-bremen.de/de/essen-trinken/uni-mensa shows prices for price group B as well. This could be served per meal as well in requests.

No tests yet, for automated updates etc. we should implement basic tests that garantuee the apis functions. Most importantly we should test for the DOM Structure of the scraped website. This should also ensure that the frontend continues to work: https://github.com/LegacyMode/Bremen-Mensa-Frontend

right now provides a

404 Not Found
nginx/1.10.3 (Ubuntu)