Feature/experimental lambert 72 support

.github/workflows/latest.yml

@@ -0,0 +1,26 @@
+name: Latest
+
+on:
+  push:
+    branches:
+      - "master"
+env:
+  CARGO_TERM_COLOR: always
+jobs:
+  docker:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+      - name: Login to DockerHub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKER_USERNAME }}
+          password: ${{ secrets.DOCKER_PASSWORD }}
+      - name: Build and push
+        uses: docker/build-push-action@v5
+        with:
+          push: true
+          tags: nbittich/mu-search:latest

Dockerfile

@@ -1,6 +1,9 @@
 FROM semtech/mu-jruby-template:3.1.0
 
 LABEL maintainer="redpencil <[email protected]>"
+
+RUN apt update && apt install gdal-bin -y
+
 # 200MB
 ENV MAXIMUM_FILE_SIZE="209715200"
 # seconds

framework/elastic_query_builder.rb

@@ -184,6 +184,26 @@ def construct_es_query_term(filter_key, value)
           terms: { field => value.split(",") }
         }
       end
+    when "geo"
+      ensure_single_field_for flag, fields do |field|
+        query = value.split(",") # just for the poc
+        lat = query[0].to_f
+        lon = query[1].to_f
+        distance = query[2]
+        {
+           bool: {
+              filter: {
+                geo_distance: {
+                  distance: distance,
+                  field => {
+                      lat: lat,
+                      lon: lon
+                  }
+                }
+              }
+            }
+        }
+      end
     when "fuzzy_phrase"
       ensure_single_field_for flag, fields do |field|
         clauses = value.split(" ").map do |word|

lib/mu_search/document_builder.rb

@@ -2,6 +2,8 @@
 
 require '/usr/src/app/sinatra_template/utils' # provided by template
 require_relative './property_definition'
+
+
 module MuSearch
   ##
   # This class is responsible for building JSON documents from an IndexDefinition
@@ -12,7 +14,7 @@ def initialize(tika:, sparql_client:, attachment_path_base:, logger:)
       @attachment_path_base = attachment_path_base
       @cache_path_base = "/cache/"
       @logger = logger
-    end
+end
 
     ##
     # Builds a document to index for the given resource URI and index_definition
@@ -122,6 +124,8 @@ def construct_document_to_index(uri: nil, definitions: property_definitions)
 
         if definition.type == "simple"
           index_value = build_simple_property(matching_values)
+        elsif definition.type == "lambert-72"
+          index_value = build_lambert_property(matching_values)
         elsif definition.type == "language-string"
           index_value = build_language_property(matching_values)
         elsif definition.type == "attachment"
@@ -138,6 +142,24 @@ def construct_document_to_index(uri: nil, definitions: property_definitions)
       Hash[key_value_tuples]
     end
 
+    def build_lambert_property(values)
+        loc_map = Hash.new {|hash, key| hash[key] = 0.0}
+        values.collect do |value|
+            ## assuming <http://www.opengis.net/def/crs/EPSG/0/31370> POINT(160167.27757517056 168249.60765740927)
+            match = value.to_s.match(/POINT\(([\d.]+)\s([\d.]+)\)/)
+            next unless match && match.length == 2
+            x_lambert = match[1].to_f  
+            y_lambert = match[2].to_f
+            next unless x_lambert && y_lambert
+            output = `echo "#{x_lambert} #{y_lambert}" | gdaltransform -s_srs EPSG:31370 -t_srs EPSG:4326`
+            lon, lat, _ = output.split(' ')
+            next unless lon && lat
+            loc_map["lon"] = lon
+            loc_map["lat"] = lat
+        end
+        [loc_map]
+    end
+
     # Get the array of values to index for a given SPARQL result set of simple values.
     # Values are constructed based on the literal datatype.
     def build_simple_property(values)

lib/mu_search/property_definition.rb

@@ -1,6 +1,6 @@
 module MuSearch
   class PropertyDefinition
-    PROPERTY_TYPES = ["simple", "nested", "attachment", "language-string"]
+    PROPERTY_TYPES = ["simple", "nested", "attachment", "language-string", "lambert-72"]
     attr_reader :name, :type, :rdf_type, :path, :pipeline, :sub_properties
 
     def initialize(name: , path:,  type: "auto", rdf_type: nil, sub_properties:)
@@ -29,8 +29,12 @@ def self.from_json_config(name, config)
             from_json_config(subname, subconfig)
           end
           rdf_type = config["rdf_type"]
-        elsif config.key?("type") && config["type"] == "language-string"
-          type = "language-string"
+        elsif config.key?("type") 
+            if config["type"] == "language-string"
+                type = "language-string"
+            elsif config["type"] == "lambert-72"
+                type = "lambert-72"
+            end
         end
       elsif config.is_a?(Array)
         path = config