From de6f1cf45fb53812036c72fca13c84afe10bcf60 Mon Sep 17 00:00:00 2001
From: Mauricio Fagundes <mauricio.fagundes@gmail.com>
Date: Tue, 15 Aug 2023 17:39:39 -0300
Subject: [PATCH] Refactor/search enhancement (#309)

---
 basedosdados_api/api/v1/search_indexes.py | 13 ++-----------
 basedosdados_api/api/v1/views.py          | 11 +++++++++--
 2 files changed, 11 insertions(+), 13 deletions(-)

diff --git a/basedosdados_api/api/v1/search_indexes.py b/basedosdados_api/api/v1/search_indexes.py
index 71aad97b..c5c2d7ae 100644
--- a/basedosdados_api/api/v1/search_indexes.py
+++ b/basedosdados_api/api/v1/search_indexes.py
@@ -14,7 +14,7 @@ class DatasetIndex(indexes.SearchIndex, indexes.Indexable):
 
     organization_id = indexes.CharField(model_attr="organization__id", null=True)
     organization_slug = indexes.CharField(model_attr="organization__slug")
-    organization_name = indexes.CharField(model_attr="organization__name")
+    organization_name = indexes.EdgeNgramField(model_attr="organization__name")
     organization_description = indexes.CharField(
         model_attr="organization__description", null=True
     )
@@ -27,7 +27,7 @@ class DatasetIndex(indexes.SearchIndex, indexes.Indexable):
 
     table_ids = indexes.MultiValueField(model_attr="tables__id", null=True)
     table_slugs = indexes.MultiValueField(model_attr="tables__slug", null=True)
-    table_names = indexes.MultiValueField(model_attr="tables__name", null=True)
+    table_names = indexes.EdgeNgramField(model_attr="tables__name", null=True)
     table_descriptions = indexes.EdgeNgramField(
         model_attr="tables__description", null=True
     )
@@ -142,15 +142,6 @@ def prepare(self, obj):
         table_ids = data.get("table_slugs", [])
         if table_ids:
             data["tables"] = []
-            for i in range(len(table_ids)):
-                data["tables"].append(
-                    {
-                        "id": data.get("table_ids", [])[i],
-                        "name": data.get("table_names", [])[i],
-                        "slug": data.get("table_slugs", [])[i],
-                        "is_closed": data.get("table_is_closed", [])[i],
-                    }
-                )
             data["total_tables"] = len(table_ids)
         else:
             data["total_tables"] = 0
diff --git a/basedosdados_api/api/v1/views.py b/basedosdados_api/api/v1/views.py
index 1ce7e95b..95d21daa 100644
--- a/basedosdados_api/api/v1/views.py
+++ b/basedosdados_api/api/v1/views.py
@@ -34,10 +34,14 @@ def get(self, request, *args, **kwargs):
 
         storage = get_storage_class()
 
-        # If query is empty, query all datasets
         if not query:
+            # If query is empty, query all datasets
             query = {"match_all": {}}
+            # Factor to multiply the number of tables by
+            # Has no effect if no query is passed
+            n_table_factor = 0
         else:
+            # If query is not empty, query datasets and tables
             query = {
                 "bool": {
                     "should": [
@@ -50,9 +54,12 @@ def get(self, request, *args, **kwargs):
                             }
                         },
                         {"match": {"name.edgengram": query}},
+                        {"match": {"table_names.edgengram": query}},
+                        {"match": {"organization_name.edgengram": query}},
                     ]
                 }
             }
+            n_table_factor = 2
 
         all_filters = []
 
@@ -147,7 +154,7 @@ def get(self, request, *args, **kwargs):
                             "field_value_factor": {
                                 "field": "n_tables",
                                 "modifier": "square",
-                                "factor": 2,
+                                "factor": n_table_factor,
                                 "missing": 0,
                             }
                         },