This provider allow Koop to fetch and transform data from an Elasticsearch instance.
From command line in the folder containing the project run:
npm install @koopjs/provider-elasticsearch --save
In your Koop instance, register the provider like:
const esProvider = require('@koopjs/provider-elasticsearch')
koop.register(esProvider)The latest version of this provider can be installed with the Koop CLI. See the Koop CLI docs for details on setting up a Koop instance and add providers with the CLI.
To suppress KoopJS warnings from the console output, run with an environment variable of KOOP_WARNINGS="suppress". In powershell, this will look like: $env:KOOP_WARNINGS="suppress" ; node main.js
Setting an environment variable of KOOP_ENV_KEYS allows saving of usernames and passwords to connect to elastic search
without having to store them in the configuration file. The format for this variable is
esId,user,password||esId2,user2,password2. The esId is the id given to the connection in the configuration file as
shown below in the Basic Config File Structure section.
npm start
By default, all services only have a single layer, as defined in the index configuration. It is possible to configure one or more sub-layers in the subLayers section of the configuration. Each of these sub layers can be a predefined sub layer or you can create your own.
The following is a sample configuration file showing most capabilities
{
"esConnections": {
"esNameForServiceURLs": {
"id": "esNameForServiceURLs",
"protocol": "http://",
"port": 9200,
"hosts": [
"localhost"
],
"shapeIndices": {
"states": {
"geometryField": "geometry",
"geometryType": "Polygon",
"joinField": "NAME"
}
},
"indices": {
"myService1": {
"index": "indexName",
"allowMultiPoint": false,
"caching": {
"enabled": true,
"seconds": 600
},
"geometryField": "geometry.coordinates",
"geometryType": "geo_point",
"returnFields": [
"lastUpdate",
"createdAt",
"name"
],
"dateFields": [
"lastUpdate",
"createdAt"
],
"idField": "OBJECTID",
"subLayers": [
{"name": "geohash"}
],
"maxResults": 1000,
"maxLayerInfoResults": 1
},
"tableService": {
"index": "indexNoShape",
"allowMultiPoint": false,
"isTable": true,
"returnFields": [
"state",
"county",
"date"
],
"dateFields": [
"date"
],
"maxResults": 1000
},
"joinService": {
"index": "indexToJoin",
"allowMultiPoint": false,
"returnFields": [
"date",
"country",
"state.name"
],
"dateFields": [
"date"
],
"subLayers": [],
"shapeIndex": {
"name": "states",
"joinField": "state.name"
},
"maxResults": 1000
},
"polyService": {
"index": "polygonIndex",
"allowMultiPoint": false,
"geometryField": "geometry",
"geometryType": "MultiPolygon",
"reversePolygons": true,
"returnFields": [
"properties.date",
"properties.count",
"properties.state_name"
],
"dateFields": [
"properties.date"
],
"subLayers": [],
"maxResults": 1000
}
}
}
}
}shapeIndices includes all indices that will be used for their shapes on other services. All fields are mandatory.indices this object has a property for every service to be created. The name of the property will be the service
name.Index Properties
index is the name of the ElasticSearch indexisTable treat this service as a table, ignoring geometryallowMultiPoint is only important for point services and allows more than one point per featurecaching allows local in-memory caching. Only use this for services that have no more documents than your maxResults settinggeometryField is the full location of the geometrygeometryType can be geo_point, Point, MultiPoint, Polyline, MultiLineString, Polygon and MultiPolygonreversePolygons if the stored polygons do not follow the right-hand rule setting this to true will fix this.
Polygons that do not follow the right-hand rule will not be displayed as feature services without setting this to true.returnFields includes all fields that will be returned by the feature servicedateFields includes any return fields that should be treated as datesidField if the index includes a field that can be treated as the OBJECTID field, this should be setsubLayers any sub-layers to be used. Provided sub layers are geohash_aggregation and geotile_aggregation but custom subLayers may be used here as well.maxResults the maximum features returned from a single requestmaxLayerInfoResults maximum number of features to return for a layer info request (Normally 1) NOTE: Do not use this on a layer with caching enabled.shapeIndex if the service will join to a shapeIndex for geometry list the name of the index (defined in
shapeIndices) and the joinField from this index.vectorLayerID the 0 based index of the subLayer to use for vector outputvectorStyle an optional map-box style to use for the vector layerAll subLayer configurations must include a name to map to a registered subLayer class. Other than that an options object contains
any other needed information.
offset which is used by both provided subLayers is the offset value passed in by esri clients. This value is in meters and is larger for higher zoom levels.aggregationFields can be used to pass in sub-aggregations to elastic search. The fields will be added to the output for display in pop-ups and/or for use in symbology.GeoHash
precision is between 1-12 for geohash.
{
"name": "geohash_aggregation",
"options": {
"tileConfig": [
{ "precision": 8, "offset": 16 },
{ "precision": 6, "offset": 10000 },
{ "precision": 4, "offset": 32000 },
{ "precision": 2, "offset": 640000 }
],
"aggregationFields": {
"speed_avg": {
"avg": { "field": "speed"}
},
"speed_min": {
"min": { "field": "speed"}
}
}
}
}GeoTile
precision is between 0-29 for geotile.
{
"name": "geotile_aggregation",
"options": {
"tileConfig": [
{ "precision": 22, "offset": 16 },
{ "precision": 16, "offset": 10000 },
{ "precision": 10, "offset": 32000 },
{ "precision": 4, "offset": 640000 }
],
"aggregationFields": {
"speed_avg": {
"avg": { "field": "speed"}
},
"speed_min": {
"min": { "field": "speed"}
}
}
}
}Support for connections to Amazon OpenSearch clusters has been added. This functionality is new and had not been thoroughly tested. Please log any issues encountered when using/testing.
The OpenSearch configurations follow the same pattern as elasticsearch connections and can even be included in the same
configuration file. Create an osConnections object sibling to the esConnections object within the configuration
file. From there the configuration is the same as with elasticsearch connections except for cluster connection
information. All OpenSearch connection information is stored within a connectConfig object. An example is given below.
{
"osConnections": {
"osCluster1": {
"id": "osCluster1",
"connectConfig": {
"node": "http://localhost:9200"
},
"indices": {
// same as with elastic configuration
}
}
}
}mapReturnValues is an object that can contain keys that are field names that in turn have their own keys equal to field
values mapped to object values. Example below. __defaultmapping is not required.
{
"mapReturnValues": {
"fieldName": {
"returnedValue1": "mappedValue",
"__defaultmapping": "defaultValue"
}
}
}IN DEVELOPMENT
mapFieldNames can be used to specify a different return field than what is specified within returnFields
{
"mapFieldNames": {
"fieldName": "mappedFieldName"
}
}