Hey all, I'm looking to get a better understanding...
# community-help
w
Hey all, I'm looking to get a better understanding of the text_match scoring I'm seeing for a query. For context I have a collection of about 10k movies and tv shows and I'm querying for "harry p" on this collection. this is what my query params look like going to typesense:
Copy code
{'filter_by': 'product:some_product&&_start_timestamp:<=1750269396&&_end_timestamp:>1750269396', 'sort_by': '', 'q': 'harry p', 'query_by': 'sort_title,language_data.sort_title,people.people.name,genres.name,genres.frenchGenre,tags,language_data.tags,short_description,language_data.short_description,description,language_data.description', 'query_by_weights': '127,126,1,0,0,126,126,1,1,1,1', 'pinned_hits': '', 'page': '1', 'per_page': '10', 'infix': 'always,always,off,off,off,off,off,off,off,off,off', 'text_match_type': 'max_weight', 'prioritize_token_position': 'true'}
Attached are h1 and h2 jsons, h1 is the first result, and h2 is the 40th. So quite a disparity, but looking at the info I don't see a vast difference between these 2, ideally both items should be on the first page. What could be the reason for the scores to be so vastly different even though the items are similar? Any help would be appreciated.
especially considering that the title is one of the highest weights
k
Have you tried disabling infix search?
w
Hey forgot to update here, the fix was to pass in the prefix param specifically, I enabled it only for title fields and nothing else and that gave the desired outcome.
Hopefully this will be helpful for people in the future 😂
🙌 1