Xponents REST: behavior with phrase tagging needs improvement; here a…

…ll taxons are reported
OpenSextant · Jun 10, 2024 · 947a78a · 947a78a
1 parent 89e4f8b
commit 947a78a
Show file tree

Hide file tree

Showing 7 changed files with 42 additions and 16 deletions.
diff --git a/RELEASE.md b/RELEASE.md
@@ -13,6 +13,9 @@ Visit https://github.com/OpenSextant/Xponents/releases for latest library releas
   were replaced.
   * Migrated to a separate repo:  Find this now maintained separately at [`Xponents-Core` repo](https://github.com/OpenSextant/Xponents-Core)
   This is the first release ./Core/ folder will not appear in this source tree.
+* **Xponents REST**: 
+  * Noise filtering;  added `minlen` parameter to Xlayer (Xponents REST) service.  By default it filters out trivial stuff. About 3-5% reduction in 
+   noise tags from codes, short names, etc, -- things typically not useful geocodings nor typically names of places.
 
 **Xponents 3.6.7**: Springy
 * **Core API**

diff --git a/pom.xml b/pom.xml
@@ -5,7 +5,7 @@
   <modelVersion>4.0.0</modelVersion>
   <groupId>org.opensextant</groupId>
   <artifactId>opensextant-xponents</artifactId>
-  <version>3.7.3</version>
+  <version>3.7.4</version>
   <packaging>jar</packaging>
   <name>OpenSextant Xponents</name>
   <description>An information extraction toolkit focused on geography and temporal entities</description>

diff --git a/src/main/java/org/opensextant/output/Transforms.java b/src/main/java/org/opensextant/output/Transforms.java
@@ -296,6 +296,17 @@ private static JsonObject populateMatch(final TextMatch m) {
         return o;
     }
 
+    private static JsonArray populateTaxon(List<Taxon> taxons) {
+        JsonArray arr = new JsonArray();
+        for (Taxon t : taxons) {
+            JsonObject tj = new JsonObject();
+            tj.put("cat", t.catalog);
+            tj.put("taxon", t.name);
+            arr.add(tj);
+        }
+        return arr;
+    }
+
     /**
      * Return seconds of epoch.
      *
@@ -349,8 +360,11 @@ public static JsonObject toJSON(final List<TextMatch> matches, final Parameters
                         // Only get one taxon from this match. That is sufficient, but not perfect.
                         Taxon n = match.getTaxons().get(0);
                         JsonObject node = populateMatch(name);
-                        node.put("taxon", n.name); // Name of taxon
+                        node.put("taxon", n.name); // Name of taxon, just the first and most common.
                         node.put("catalog", n.catalog); // Name of catalog or source
+                        if (match.getTaxons().size() > 1) {
+                            node.put("all-taxons", populateTaxon(match.getTaxons())); // Name of taxon, just the first and most common.
+                        }
                         node.put("method", "TaxonMatcher");
                         resultArray.add(node);
                     }

diff --git a/src/main/java/org/opensextant/xlayer/server/TaggerResource.java b/src/main/java/org/opensextant/xlayer/server/TaggerResource.java
@@ -128,7 +128,7 @@ protected void parseParameters(Parameters p, Set<String> kv) {
         }
 
         // Request tagging on demand.
-        p.tag_all_taxons = kv.contains("all-taxons");
+        p.tag_all_taxons = kv.contains("all_taxons");
         p.tag_taxons = (kv.contains("taxons") || kv.contains("orgs") || kv.contains("persons"));
         p.tag_patterns = kv.contains("patterns") || kv.contains("dates");
 

diff --git a/src/main/java/org/opensextant/xlayer/server/xgeo/XlayerRestlet.java b/src/main/java/org/opensextant/xlayer/server/xgeo/XlayerRestlet.java
@@ -7,6 +7,7 @@
 import org.opensextant.extraction.MatchFilter;
 import org.opensextant.extractors.geo.PlaceGeocoder;
 import org.opensextant.extractors.geo.PostalGeocoder;
+import org.opensextant.extractors.xtax.TaxonMatcher;
 import org.opensextant.extractors.xtemporal.XTemporal;
 import org.opensextant.processing.Parameters;
 import org.opensextant.xlayer.server.XlayerApp;
@@ -39,9 +40,10 @@ public synchronized Restlet createInboundRoot() {
             banner();
             configure();
             Context ctx = getContext();
-            ctx.getAttributes().put("xgeo", tagger);
-            ctx.getAttributes().put("xtemp", dateTagger);
-            ctx.getAttributes().put("xpostal", postalGeocoder);
+            ctx.getAttributes().put(XponentsGeotagger.GEO_TAGGER, tagger);
+            ctx.getAttributes().put(XponentsGeotagger.DATE_TAGGER, dateTagger);
+            ctx.getAttributes().put(XponentsGeotagger.POSTAL_TAGGER, postalGeocoder);
+            ctx.getAttributes().put(XponentsGeotagger.TAXON_TAGGER, phraseTagger);
 
             ctx.getAttributes().put("version", this.version);
             info("%%%%   Xponents Geo Phase Configured");
@@ -59,6 +61,7 @@ public synchronized Restlet createInboundRoot() {
     private PlaceGeocoder tagger = null;
     private XTemporal dateTagger = null;
     private PostalGeocoder postalGeocoder = null;
+    private TaxonMatcher phraseTagger = null;
 
     /**
      * @throws ConfigException
@@ -73,6 +76,8 @@ public void configure() throws ConfigException {
         tagger.setParameters(taggerParams);
         tagger.configure();
 
+        phraseTagger = new TaxonMatcher();
+
         // TODO: refine this filter list. Use "/filters/non-placenames,user.csv" going forward.
         //
         String userFilterPath = "/filters/non-placenames,user.csv";

diff --git a/src/main/java/org/opensextant/xlayer/server/xgeo/XponentsGeotagger.java b/src/main/java/org/opensextant/xlayer/server/xgeo/XponentsGeotagger.java
@@ -12,6 +12,7 @@
 import org.opensextant.extractors.geo.PlaceCandidate;
 import org.opensextant.extractors.geo.PlaceGeocoder;
 import org.opensextant.extractors.geo.PostalGeocoder;
+import org.opensextant.extractors.xtax.TaxonMatcher;
 import org.opensextant.extractors.xtemporal.XTemporal;
 import org.opensextant.output.Transforms;
 import org.opensextant.processing.Parameters;
@@ -157,25 +158,28 @@ public Representation process(TextInput input, Parameters jobParams) {
             List<TextMatch> matches = new ArrayList<>();
 
             // BOTH geo and taxons could be requested:  features = "geo", "all-taxons"
-            // `geo` tagging
-            if (tag_geo(jobParams) || tag_taxons(jobParams)) {
+            if (!tag_geo(jobParams) && tag_taxons(jobParams)) {
+                // Taxonomic tags only
+                TaxonMatcher phraseTagger = (TaxonMatcher) getExtractor(TAXON_TAGGER);
+                matches.addAll(phraseTagger.extract(input, jobParams));
+            } else if (tag_geo(jobParams) || tag_taxons(jobParams)) {
+                // Geotagging
                 PlaceGeocoder xgeo = (PlaceGeocoder) getExtractor(GEO_TAGGER);
                 matches.addAll(xgeo.extract(input, jobParams));
             }
+
             if (jobParams.tag_patterns) {
                 XTemporal xt = (XTemporal) getExtractor(DATE_TAGGER);
                 matches.addAll(xt.extract(input));
             }
             if (jobParams.tag_postal) {
                 PostalGeocoder pg = (PostalGeocoder) getExtractor(POSTAL_TAGGER);
-                if (pg != null) {
-                    if (tag_geo(jobParams)) {
-                        // OPTIMIZATION: reuse matches accumulated so far to prevent
-                        // PostalGeocoder from repeating extract()
-                        pg.setGeneralMatches(matches);
-                    }
-                    matches.addAll(pg.extract(input));
+                if (tag_geo(jobParams)) {
+                    // OPTIMIZATION: reuse matches accumulated so far to prevent
+                    // PostalGeocoder from repeating extract()
+                    pg.setGeneralMatches(matches);
                 }
+                matches.addAll(pg.extract(input));
             }
             if (isDebug()) {
                 debug(String.format("CURRENT MEM USAGE(K)=%d", RuntimeTools.reportMemory()));

diff --git a/src/main/resources/banner.txt b/src/main/resources/banner.txt
@@ -12,5 +12,5 @@
 //                       
 //                   Xponents Geotagger version 3.7 (2024-Q1)
 //  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
-VERSION: 3.7.3
+VERSION: 3.7.4