perf(csv_export): resolve timeout when exporting related data

django_forest/resources/associations/views/csv.py

@@ -1,5 +1,6 @@
 import logging
 import csv
+import time
 
 from django_forest.resources.associations.utils import AssociationView
 from django_forest.resources.utils.csv import CsvMixin
@@ -13,6 +14,7 @@
 
 class CsvView(SmartFieldMixin, JsonApiSerializerMixin, CsvMixin, AssociationView):
     def get(self, request, pk, association_resource):
+        t0 = time.time()
         try:
             association_field = get_association_field(self.Model, association_resource)
         except Exception as e:
@@ -25,15 +27,25 @@ def get(self, request, pk, association_resource):
             queryset = getattr(self.Model.objects.get(pk=pk), association_resource).all()
 
             params = request.GET.dict()
+
+            t2 = time.time()
             # enhance queryset
             queryset = self.enhance_queryset(queryset, RelatedModel, params, request, apply_pagination=False)
+            for _ in queryset[0:1]:  # force SQL request execution
+                pass
 
+            t3 = time.time()
+            logger.warning(f"-- timing enhance_queryset: {(t3 - t2):.2f} seconds")
             # handle smart fields
             self.handle_smart_fields(queryset, RelatedModel._meta.db_table, parse_qs(params), many=True)
 
+            t4 = time.time()
+            logger.warning(f"-- timing smart fields: {(t4 - t3):.2f} seconds")
             # json api serializer
             data = self.serialize(queryset, RelatedModel, params)
 
+            t5 = time.time()
+            logger.warning(f"-- timing json api serialization: {(t5 - t4):.2f} seconds")
             response = self.csv_response(params['filename'])
 
             field_names_requested = [x for x in params[f'fields[{RelatedModel._meta.db_table}]'].split(',')]
@@ -42,4 +54,7 @@ def get(self, request, pk, association_resource):
             writer = csv.DictWriter(response, fieldnames=field_names_requested)
             writer.writerow(dict(zip(field_names_requested, csv_header)))
             self.fill_csv(data, writer, params)
+            t6 = time.time()
+            logger.warning(f"-- timing csv serialization into response body: {(t6 - t5):.2f} seconds")
+            logger.warning(f"--- total timing: {(t6 - t0):.2f} seconds")
             return response

django_forest/resources/utils/queryset/__init__.py

@@ -1,10 +1,13 @@
+import re
+
 from .filters import FiltersMixin
 from .limit_fields import LimitFieldsMixin
 from .pagination import PaginationMixin
 from .scope import ScopeMixin
 from .search import SearchMixin
 from .segment import SegmentMixin
 from django_forest.resources.utils.decorators import DecoratorsMixin
+from django_forest.utils.schema import Schema
 
 
 class QuerysetMixin(
@@ -26,7 +29,33 @@ def filter_queryset(self, queryset, Model, params, request):
                 queryset = queryset.filter(method(params, Model))
         return queryset
 
+    def join_relations(self, queryset, Model, params, request):
+        select_related = set()
+
+        collection = Schema.get_collection(Model._meta.db_table)
+        relations = [
+            field['field']
+            for field in collection["fields"]
+            if field["relationship"] is not None and field["relationship"] in ["BelongsTo", "HasOne"]
+        ]
+
+        # scope
+
+        # filters
+
+        # projection
+        for key, value in params.items():
+            if re.search(r"fields\[[^\]]+\]", key):
+                fields_for = key.split("fields[")[1][:-1]
+                if fields_for in relations:
+                    select_related.add(fields_for)
+
+        return queryset.select_related(*select_related)
+
     def enhance_queryset(self, queryset, Model, params, request, apply_pagination=True):
+        # perform inner join
+        queryset = self.join_relations(queryset, Model, params, request)
+
         # scopes + filter + search
         queryset = self.filter_queryset(queryset, Model, params, request)