Menu extracting - nlp

Question

I am interesting in extracting and structuring information about restaurant menus. What is needed is to extract the items from the menu in form category / name / price
For instance, we have the following website. Here we have a drinks sections, and there a number of items. For that website I'd like to be able to extract
Drink / Cappuccino / € 1,50
SANDWICHES / filled sandwich, pistolet (round roll) or emperor roll / € 1,30
etc ...
Of course it shouldn't be limited only to this website.
The only way I can see to handle that is applying a bunch of regexps, but I don't believe listing all possible dish names is feasible.
I know that the topic might be too broad for a question, but anyway any suggestions or references to relevant articles or books will be much appreciated.

Answer 1

This seems quite possible. You many not be able to list all possible dishes but you can list all possible categories.
Assuming that in every menu, dish names follows category name and it is followed by the price, you can identify dish names.
The algorithm will look like this:
foreach(category: category_list):
foreach(word:document):
if(category == word):
dish = Read next(if data is structures with table read next row or col)
price = Read next and check it format to see if its Currency or a price
The point is you will need to analyse different websites to understand how the information is structured and prepare your algorithm to deal with all possible structures.