Vertical categories matching enhancements (offer names integration)

api/src/main/java/org/open4goods/api/services/aggregation/services/realtime/TaxonomyRealTimeAggregationService.java

@@ -2,8 +2,10 @@
 
 import java.util.ArrayList;
 import java.util.HashMap;
+import java.util.HashSet;
 import java.util.List;
 import java.util.Map;
+import java.util.Set;
 
 import org.apache.commons.lang3.StringUtils;
 import org.open4goods.api.services.aggregation.AbstractAggregationService;
@@ -87,6 +89,42 @@ public void onProduct(Product data, VerticalConfig vConf) throws AggregationSkip
 			data.setVertical(null);
 		}
 
+
+		/////////////////////////////////////
+		// A hard filtering on datasource categories, that must at least contains one of the
+		// vertical natural names
+		////////////////////////////////////
+
+		// Building offer name bags
+		StringBuilder productNameBag = new StringBuilder();
+		data.getOfferNames().forEach(name -> {
+			productNameBag.append(StringUtils.stripAccents(name).toLowerCase());
+		});
+		String pNames = productNameBag.toString();
+
+		// TODO : Singularize
+		// TODO : Cache
+		Set<String> verticalNames = vConf.getTokenNames(taxonomyService.byId(vConf.getGoogleTaxonomyId()).getGoogleNames().values().stream().map(e->StringUtils.stripAccents(e.toLowerCase())).toList() );
+
+
+		for (String term : verticalNames) {
+
+			if (pNames.contains(term)) {
+				dedicatedLogger.info("Vertical {} confirmed by product names match for {}", vConf,data);
+				data.setVertical(vConf.getId());
+				break;
+			} else {
+				dedicatedLogger.info("Vertical {} failed on product names match, unsetting vertical for {}", vConf,data);
+				data.setVertical(null);
+			}
+
+		}
+
+
+
+
+
+
 //		// Setting no vertical if no category
 		if (data.getDatasourceCategories().size() == 0) {
 			dedicatedLogger.info("No category in {}, removing vertical", data);

commons/src/main/java/org/open4goods/commons/config/yml/ui/VerticalConfig.java

@@ -1,6 +1,7 @@
 package org.open4goods.commons.config.yml.ui;
 
 import java.util.ArrayList;
+import java.util.Collection;
 import java.util.HashMap;
 import java.util.HashSet;
 import java.util.List;
@@ -12,6 +13,7 @@
 import java.util.Set;
 import java.util.stream.Collectors;
 
+import org.apache.commons.lang3.StringUtils;
 import org.open4goods.commons.config.yml.CommentsAggregationConfig;
 import org.open4goods.commons.config.yml.attributes.AttributeConfig;
 import org.open4goods.commons.helper.IdHelper;
@@ -230,6 +232,10 @@ public class VerticalConfig{
 	 */
 	private List<FeatureGroup> featureGroups = new ArrayList<>();
 
+
+	// A local cache for token names
+	private Set<String> cacheTokenNames;
+
 
 	@Override
 	public String toString() {
@@ -250,6 +256,50 @@ public ProductI18nElements i18n(String lang) {
 		return i18n.getOrDefault(lang, i18n.get("default"));		
 	}
 
+
+	/**
+	 * Compute the token names, used to do a product.offernames matching categories
+	 * 
+	 * @param additionalNames
+	 * @return
+	 */
+	public Set<String> getTokenNames(Collection<String> additionalNames) {
+
+		if (null == cacheTokenNames) {
+
+			Set<String> verticalNames = new HashSet<String>();
+			verticalNames.add(getId());
+			verticalNames.addAll(additionalNames);
+
+			getI18n().entrySet().forEach(e -> {
+				// verticalNames.add(e.getValue().getH1Title().getPrefix().toLowerCase());
+				verticalNames.add(e.getValue().getVerticalHomeUrl());
+				// verticalNames.add(e.getValue().getUrl().getPrefix());
+
+			});
+
+			// Adding singular derivativs, removing -
+			Set<String> derivativs = new HashSet<String>();
+			verticalNames.forEach(e -> {
+				String derivativ = e.replace("-", " ");
+				derivativs.add(derivativ);
+
+				if (derivativ.endsWith("s")) {
+					derivativ = derivativ.substring(0, derivativ.length() - 1);
+				}
+				derivativs.add(derivativ);
+
+				// If composed word (laves linges"
+				derivativ = derivativ.replace("s ", " ");
+				derivativs.add(derivativ);
+
+			});
+
+			verticalNames.addAll(derivativs);
+			cacheTokenNames = verticalNames;
+		}
+		return cacheTokenNames;
+	}
 	/**
 	 * Return the participation in percentage of a score in the ecoscore
 	 * @param scoreName
@@ -767,6 +817,8 @@ public void setGenerationExcludedFromAttributesMatching(Set<String> generationEx
 	}
 
 
+
+
 
 
 

commons/src/main/java/org/open4goods/commons/services/VerticalsConfigService.java

@@ -259,8 +259,6 @@ public VerticalConfig getVerticalForCategories(Map<String, String> productCatego
 		    }
 		}
 
-
-
 		return vc;
 	}
 

verticals/src/main/resources/verticals/_default.yml

@@ -494,7 +494,7 @@ attributesConfig:
     synonyms:
       all:
       - "QTE DE PORTS HDMI"
-      - "QUANTITE DE PORTS HDMI"
+      - "QUANTITE DE PORTS HDMI"      
 
     parser:
       normalize: true
@@ -557,7 +557,7 @@ attributesConfig:
       default: "Weight"
       fr: "Poids"
     filteringType: "NUMERIC"
-    asScore: false
+    asScore: true
 
     synonyms:
       all: