jmankhan / bergdining Goto Github PK

Pseudocode is designed

Fill with dummy data for now

Note that this is a Maven project, so add dependencies through Maven

Cache new menus within first 2 hours due to expected large download requests.
Save to database after and pull from it rather than connecting to website and re-parsing

Use local postgres db

Fill with a link to the wiki, perhaps.
This server is intended to be used by a computer, not people. The homepage does not need to be pretty

Done. "Dictionary" text file was created with keywords. Incoming menu item names were compared to dictionary and image ids were assigned based on matching keywords.

e.g. "Grilled Turkey Burger" was assigned the "Burger" id. Priority is given to broader categories of food. The user would expect a picture of a burger rather than a turkey, even though the item is both. Priorities will probably need to be adjusted as more items are analyzed. Priority is assigned by place in list (higher on list = higher priority)

Not all items are able to be categorized, such as those for which we have no image or are given strange or broad names (e.g. "Au jus", what is that?). This is expected to be rectified with time as more and more menu items are analyzed.

Discuss access to master dining list and parsing options
Find a way to get more images

Have a document ready by end of week describing CRUD interface

Gives a basic overview of how this server works, what heroku is, and the key components

Call the parsing method every week using Spark. We don't want to run a process continuously for that long, so I will try to find a more efficient way through the framework

Don't know if this is even necessary, but compressing it may help

Update wiki accordingly

should be a property with every menuitem