Create Unit Tests

[lgrymaloski]

Create Unit Tests to verify the functionality of the search and recommender features.

[haleyrutten]

Suggested Tests:

add and subtract as you wish, this is just a starting point

Search

• [ ] no results found
• [ ] sparse vector retrieval works on listings
• [ ] sparse vector retrieval works on users (if search is "john", results = john, john_wick, little_john)
• [ ] sparse vector retrieval works for multiple words (ex: yellow+bike gives yellow objects and bike objects. Is a yellow bike at the top?)
• [ ] empty search sent ie. "" (if frontend allows this case)

Recommendations

• [ ] userId not found
• [ ] userId has no previous search history (i.e. cold start)
• [ ] userId has search history that matches no listings
• [ ] userId has search history that matches some listings
• [ ] userId has search history with multiple words
• [ ] userId has search history with multiple searches for same thing (ex: "bike", "yellow+bike", "green+bike","lamp") -> ensure each listing is only recommended once, and bikes are more recommended.
• [ ] user can ignore a listing. (are all listings of same type ignored, or just the singular listing?)

higher level tests

• [ ] check that searches are added to search history
• [ ] check do not recommend persists for a user
• [ ] check do not show charity persists for user
• [ ] ensure recommendations updated as user searches for new things.
• [ ] (requested by TA) test case for re-indexing: "you try to edit a listing to match to your keywords, after re-indexing, that listing is expected to shown in the search results."