ainemitch
commented
3 years ago Hi Julien,
Many thanks for your comments they are very helpful.
In answer to your question on the reason for wanting the duplicate values in name.default and name.${lang}, we are actually attempting to build a reverse geo-coding application which supports multiple languages and we are using requests similar to the below for retrieving the location data.
curl --location --request GET 'localhost:9200/pelias_world_1.0/_search' --header 'Content-Type: application/json' --data-raw '{
"query":{
"bool": {
"must": {
"match_all": {}
},
"filter": {
"geo_shape": {
"polygon": {
"shape":
{
"type": "point",
"coordinates" : [5.420364, 43.291582]
},
"relation": "contains"
}
}
}
}
}
}'
What we found that with the duplicate names filtered out in some cases the translated value returned was the variant name rather than the preferred name (in a comparison with a WhosOnFirst Spelunker search, for example for the locality Marseille
Without duplicates the French translation is 13000 (variant)
"bounding_box":"{\"min_lat\":43.1726492325,\"max_lat\":43.3910172471,\"min_lon\":5.22826669352,\"max_lon\":5.53250260623}", "name":{ "default":[ "Marseille", "MRS" ], .....
"fr":"13000",With the duplicates included we would get both the preferred and variant names which is what we want for our application.
"fr":[
"Marseille",
"13000"
],We will take a look at Pelias-conf and the OpenAddresses importer, thanks again for the information on those. I will close this pull request so and raise a more targeted ones with just the shape update. Thanks you again for taking the time to comment.
GEO-46: Pelias WOF data importer improvement to include geo-shape data into Pelias ES index - patch applied, tests updated and new task added to generate additional error log.