Merge branch 'sort-volumes' into 'master'

Sort search results by volume names

See merge request !120

arkindex/documents/api/search.py

@@ -19,7 +19,7 @@ class PageSearch(SearchAPIView):
     serializer_class = PageSearchResultSerializer
     template_path = 'elastic/search_nested.json'
     es_source = False
-    es_sort = ["_score", ]
+    es_sort = ["volumes", "_score", ]
     es_index = settings.ES_INDEX_PAGES
     es_type = Page.INDEX_TYPE
 
@@ -34,7 +34,7 @@ class ActSearch(SearchAPIView):
     serializer_class = ActSearchResultSerializer
     template_path = 'elastic/search_nested.json'
     es_source = False
-    es_sort = ['_score', ]
+    es_sort = ["volumes", "_score", ]
     es_index = settings.ES_INDEX_ACTS
     es_type = Act.INDEX_TYPE
 

arkindex/documents/indexer.py

@@ -27,6 +27,7 @@ class Indexer(object):
                         Act.INDEX_TYPE: {
                             "properties": {
                                 "corpus": {"type": "keyword"},
+                                "volumes": {"type": "keyword"},
                                 "transcriptions": {
                                     "type": "nested",
                                     "properties": {
@@ -51,6 +52,7 @@ class Indexer(object):
                         Page.INDEX_TYPE: {
                             "properties": {
                                 "corpus": {"type": "keyword"},
+                                "volumes": {"type": "keyword"},
                                 "transcriptions": {
                                     "type": "nested",
                                     "properties": {

arkindex/documents/models.py

@@ -385,6 +385,7 @@ class Page(Element):
         """
         return {
             'corpus': self.corpus_id,
+            'volumes': [v.name for v in Element.objects.get_ascending(self.id, type=ElementType.Volume)],
             'transcriptions': [
                 {
                     'id': t.id,
@@ -451,6 +452,7 @@ class Act(Element):
         ]
         return {
             'corpus': self.corpus_id,
+            'volumes': [v.name for v in Element.objects.get_ascending(self.id, type=ElementType.Volume)],
             'transcriptions': [
                 {
                     'id': t.id,

arkindex/documents/search.py

@@ -55,7 +55,6 @@ def search_nested_post(model, data):
 
     results = data['hits']['hits']
     elt_ids = [uuid.UUID(r['_id']) for r in results]
-    elt_scores = {uuid.UUID(r['_id']): r['_score'] for r in results}
     if not elt_ids:
         return
     tr_ids = [
@@ -73,7 +72,6 @@ def search_nested_post(model, data):
         for t in Transcription.objects.filter(id__in=tr_ids).prefetch_related('zone__image__server')
     }
 
-    elts = model.objects.filter(id__in=elt_ids).prefetch_related('corpus')
     elts_tr_ids = {
         uuid.UUID(result['_id']): [
             uuid.UUID(hit['_source']['id'])
@@ -81,11 +79,15 @@ def search_nested_post(model, data):
         ] for result in results
     }
 
-    all_paths = Element.objects.get_ascendings_paths(*elt_ids)
+    elts = list(model.objects.filter(id__in=elt_ids).prefetch_related('corpus'))
+    # Preserve the ordering given by ElasticSearch
+    ordered_elts = list(filter(None, map(lambda eid: next((e for e in elts if e.id == eid), None), elt_ids)))
 
-    for elt in elts:
+    all_paths = Element.objects.get_ascendings_paths(*(e.id for e in ordered_elts))
+
+    for elt in ordered_elts:
         elt.transcriptions_results = list(filter(None, [transcriptions.get(tid) for tid in elts_tr_ids[elt.id]]))
         elt.total_transcriptions = tr_totals[elt.id]
         elt.parent_paths = all_paths.get(elt.id, [])
 
-    return sorted(elts, key=lambda e: elt_scores[e.id], reverse=True)
+    return ordered_elts