Completion suggester fuzziness ranks typos over exact matches


(John Harry) #1

Hi,
I am using the completion suggester for predictive search with fuzziness enabled and am finding that the ranking is favouring typos over exact matches. I.e. in Kiban I use this
GET /index/typeahead/_search
{
"_source": ["suggest"],
"size": 10,
"suggest":
{
"suggestions": {
"prefix": "meat ",
"completion": {
"field": "suggest",
"size": 10,
"fuzzy": {
"min_length": 2,
"prefix_length": 1,
"fuzziness": "AUTO"
}
}
}
}
}

and get back results like meadow etc.
while removing fuzziness completeley like this:
GET /index/typeahead/_search
{
"_source": ["suggest"],
"size": 7,
"suggest":
{
"suggestions" :
{
"prefix" : "meat ",
"completion" :
{
"field" : "suggest"
}
}
}
}

returns results with the word meat in them.

some help to get word matches to be ranked higher than typos would be appreaciated.


(system) #2

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.