C4-301 fix nested stats aggs

dbmi-bgm · Sep 30, 2020 · ed51aad · ed51aad
1 parent f131284
commit ed51aad
Show file tree

Hide file tree

Showing 2 changed files with 44 additions and 11 deletions.
diff --git a/src/encoded/search/lucene_builder.py b/src/encoded/search/lucene_builder.py
@@ -729,7 +729,7 @@ def fix_nested_aggregations(search, es_mapping):
         """
         aggs_ptr = search.aggs['all_items']
         for agg in aggs_ptr:
-            if NESTED in agg:
+            if NESTED in agg and 'stats' not in agg:  # stats aggs are already correct
                 (search.aggs['all_items'][agg]  # create a sub-bucket, preserving the boolean qualifiers
                  .bucket('primary_agg',
                          'nested', path=find_nested_path(aggs_ptr.aggs[agg]['primary_agg'].field, es_mapping))
@@ -778,17 +778,38 @@ def build_facets(cls, search, facets, search_filters, string_query, request, doc
                             facet["number_step"] = "any"
                 facet_filters = cls.generate_filters_for_terms_agg_from_search_filters(query_field, search_filters,
                                                                                        string_query)
+                # stats aggregations could be nested too
+                if nested_path:
+                    facet['aggregation_type'] = 'nested:stats'
+                    aggs[facet['aggregation_type'] + ':' + agg_name] = {
+                        AGGS: {
+                            'primary_agg': {
+                                NESTED: {
+                                    PATH: nested_path
+                                },
+                                AGGS: {
+                                    'primary_agg': {
+                                        'stats': {
+                                            'field': query_field
+                                        }
+                                    }
+                                }
+                            }
+                        },
+                        FILTER: {BOOL: facet_filters}
+                    }
 
-                aggs[facet['aggregation_type'] + ":" + agg_name] = {
-                    AGGS: {
-                        'primary_agg': {
-                            'stats': {
-                                'field': query_field
+                else:
+                    aggs[facet['aggregation_type'] + ":" + agg_name] = {
+                        AGGS: {
+                            'primary_agg': {
+                                'stats': {
+                                    'field': query_field
+                                }
                             }
-                        }
-                    },
-                    FILTER: {BOOL: facet_filters}
-                }
+                        },
+                        FILTER: {BOOL: facet_filters}
+                    }
 
             else:
                 if nested_path:

diff --git a/src/encoded/search/search.py b/src/encoded/search/search.py
@@ -801,6 +801,13 @@ def format_facets(self, es_results):
                     # Used for fields on which can do range filter on, to provide min + max bounds
                     for k in aggregations[full_agg_name]['primary_agg'].keys():
                         result_facet[k] = aggregations[full_agg_name]['primary_agg'][k]
+
+                # nested stats aggregations have a second "layer" for reverse_nested
+                elif facet['aggregation_type'] == 'nested:stats':
+                    result_facet['total'] = aggregations[full_agg_name]['primary_agg']['doc_count']
+                    for k in aggregations[full_agg_name]['primary_agg']['primary_agg'].keys():
+                        result_facet[k] = aggregations[full_agg_name]['primary_agg']['primary_agg'][k]
+
                 else:  # 'terms' assumed.
 
                     # Shift the bucket location
@@ -831,10 +838,15 @@ def format_facets(self, es_results):
 
                 if len(aggregations[full_agg_name].keys()) > 2:
                     result_facet['extra_aggs'] = {k: v for k, v in aggregations[full_agg_name].items() if
-                                                  k not in ('doc_count', "primary_agg")}
+                                                  k not in ('doc_count', 'primary_agg')}
 
             result.append(result_facet)
 
+        # TODO ALEX: Client will reject 'nested:stats' so overwritten here
+        for facet in result:
+            for k, v in facet.items():
+                if k == 'aggregation_type' and v == 'nested:stats':
+                    facet[k] = 'stats'
         return result
 
     @staticmethod