Filter repatriated datasets against country datasets (do not use repa…

…tried flag in query)
biodiversitydata-se · Jul 11, 2024 · 3cec4d6 · 3cec4d6
1 parent c51d983
commit 3cec4d6
Showing 1 changed file with 13 additions and 2 deletions.
diff --git a/src/main/groovy/au/org/ala/collectory/resources/gbif/GbifRepatDataSourceAdapter.groovy b/src/main/groovy/au/org/ala/collectory/resources/gbif/GbifRepatDataSourceAdapter.groovy
@@ -17,16 +17,24 @@ class GbifRepatDataSourceAdapter extends GbifDataSourceAdapter {
     static final SOURCE = "GBIF_REPATRIATION"
     GbifService gbifService
 
-    static final String OCCURRENCE_REPAT_SEARCH = "occurrence/search?repatriated=true&country={0}&type={1}&offset=0&limit=0&facet=datasetKey&facetLimit=10000"
+    static final String OCCURRENCE_REPAT_SEARCH = "occurrence/search?country={0}&type={1}&offset=0&limit=0&facet=datasetKey&facetLimit=10000"
+    static final String COUNTRY_DATASETS = "dataset?country={0}&limit=10000"
 
     GbifRepatDataSourceAdapter(DataSourceConfiguration configuration) {
         super(configuration)
      }
 
+    List<String> fetchCountryDatasetKeys() {
+        String url = MessageFormat.format(COUNTRY_DATASETS, configuration.country)
+        JSONObject json = getJSONWS(url, false)
+        json.results.collect{ it -> it.key }
+    }
+
     @Override
     List<Map> datasets() throws ExternalResourceException {
         def keys = []
         def datasets = []
+        def countryDatasetKeys = fetchCountryDatasetKeys()
 
         LOGGER.info("Requesting dataset lists configuration.country: ${configuration.country}")
         String url = MessageFormat.format(OCCURRENCE_REPAT_SEARCH, configuration.country, configuration.recordType)
@@ -35,7 +43,10 @@ class GbifRepatDataSourceAdapter extends GbifDataSourceAdapter {
             json.facets[0].counts.each {
                 keys << it.name
 
-                if (it.count >= configuration.minRecordCount && it.count <= configuration.maxRecordCount && datasets.size() < configuration.maxNoOfDatasets) {
+                if (!countryDatasetKeys.contains(it.name) &&
+                        it.count >= configuration.minRecordCount &&
+                        it.count <= configuration.maxRecordCount &&
+                        datasets.size() < configuration.maxNoOfDatasets) {
                     LOGGER.info("Getting metadata for ${it.name}  = ${it.count}")
                     def dataset = getDataset(it.name, it.count)
                     if (dataset.name) {