Closed fpurcell closed 5 years ago
Here, the search service weights multiple (fallback) records with the same weight, but the sorting is not clear why a point farther away appears higher in the response:
PELIAS QUERY: https://ws-st.trimet.org/pelias/v1/search?text=888%20SE%20Main%20St
NOTE: weighting 'interpolated' hits higher than 'fallback' hits is perfectly reasonable, despite the distance. So the following result is preferred.
PELIAS QUERY: https://ws-st.trimet.org/pelias/v1/search?text=822%20Main%20St
https://ws-st.trimet.org/pelias/v1/search?text=1800%20SE%20MLK%20Blvd
A handful of random addresses that basically match '1800 SE .* Blvd' come back via search with the same confidence value, but the sorting based on distance didn't happen.
https://ws-st.trimet.org/pelias/v1/autocomplete?text=Fred%20Meyer%20HOP doesn't even try on distances, I've found.
Still Funky (when filtering on stops layer):
https://ws-st.trimet.org/pelias/v1/search?text=834%20SE%20Lambert&sources=transit&layers=stops
going to close for now, since in practical use sorting, feels pretty good --- will reopen if need be from future user feedback
result sorting
I want Pelias to sort results by distance and match percentage attributes
I want Pelias to have the same 'distance' and 'confidence' attributes that I see in search also in autocomplete.