Skip to content

Search Match and Weight Testing

Ruth edited this page Oct 26, 2022 · 4 revisions

The challenge in creating search match and weight testing is that the catalog data changes all the time. The tests recommended below all operate based on results which "contain" (should be case-insensitive) vs. match and results numbers "fewer than X" where X is higher than the actual expected matches but lower than the matches from the initial Catalog launch.

Term Field Expected Results Expected Result Count
Because of Winn-Dixie Title all results title_tsim contains "Because of Winn-Dixie" fewer than 10
Because of Winn-Dixie Keyword first 6 results title_tsim all contain "Because of Winn-Dixie" fewer than 15
pittsburgh sanborn fire insurance map Keyword first 7 results title_tsim contains words "insurance" "maps" "pittsburgh" and "pennsylvania" in the titles. (this is not what we're searching for but it's the results we expect to get up front) fewer than 50
ann cooper albright Keyword first 10 results either author_tsim or author_addl_tsim contains "Albright, Ann Cooper" fewer than 50
Phenomenological Gateway to Reality Keyword none 0 results
"john earl haynes" Keyword first 9 results either author_tsim or author_addl_tsim should contain "Haynes, John Earl" fewer than 50
american journal of psychoanalysis Keyword first 2 results title_tsim should contain "american journal of psychoanalysis" fewer than 20
failed empire zubok Keyword - fewer than 5
Karl Case, Ray Fair Keyword first 5 results title_tsim should contain "principles" fewer than 25
Scientific Integrity, Francis Macrina Keyword first 5 results title_tsim should contain "scientific integrity" fewer than 10
Embodied Food Politics Keyword first result title_tsim should contain "embodied food politics" fewer than 50
Clone this wiki locally