Allowing multiple load and apply statements

Author: petyaslavova

docker-compose.yml

@@ -1,7 +1,7 @@
 ---
 # image tag 8.0-RC2-pre is the one matching the 8.0 GA release
 x-client-libs-stack-image: &client-libs-stack-image
-  image: "redislabs/client-libs-test:${CLIENT_LIBS_TEST_STACK_IMAGE_TAG:-8.2}"
+  image: "redislabs/client-libs-test:${CLIENT_LIBS_TEST_STACK_IMAGE_TAG:-8.4-RC1-pre.2}"
 
 x-client-libs-image: &client-libs-image
   image: "redislabs/client-libs-test:${CLIENT_LIBS_TEST_IMAGE_TAG:-8.4-RC1-pre.2}"

redis/commands/search/hybrid_query.py

@@ -228,18 +228,21 @@ def __init__(self) -> None:
         """
         Create a new hybrid post processing configuration object.
         """
-        self._load_fields = []
-        self._groupby = []
-        self._apply = []
+        self._load_statements = []
+        self._apply_statements = []
+        self._groupby_statements = []
         self._sortby_fields = []
         self._filter = None
         self._limit = None
 
     def load(self, *fields: str) -> Self:
         """
-        Add load parameters to the query.
+        Add load statement parameters to the query.
         """
-        self._load_fields = fields
+        if fields:
+            fields_str = " ".join(fields)
+            fields_list = fields_str.split(" ")
+            self._load_statements.extend(("LOAD", len(fields_list), *fields_list))
         return self
 
     def group_by(self, fields: List[str], *reducers: Reducer) -> Self:
@@ -262,7 +265,7 @@ def group_by(self, fields: List[str], *reducers: Reducer) -> Self:
             if reducer._alias is not None:
                 ret.extend(("AS", reducer._alias))
 
-        self._groupby.extend(ret)
+        self._groupby_statements.extend(ret)
         return self
 
     def apply(self, **kwexpr) -> Self:
@@ -274,11 +277,14 @@ def apply(self, **kwexpr) -> Self:
                 the alias for the projection, and the value is the projection
                 expression itself, for example `apply(square_root="sqrt(@foo)")`.
         """
+        apply_args = []
         for alias, expr in kwexpr.items():
             ret = ["APPLY", expr]
             if alias is not None:
                 ret.extend(("AS", alias))
-            self._apply.extend(ret)
+            apply_args.extend(ret)
+
+        self._apply_statements.extend(apply_args)
 
         return self
 
@@ -310,14 +316,12 @@ def limit(self, offset: int, num: int) -> Self:
 
     def build_args(self) -> List[str]:
         args = []
-        if self._load_fields:
-            fields_str = " ".join(self._load_fields)
-            fields = fields_str.split(" ")
-            args.extend(("LOAD", len(fields), *fields))
-        if self._groupby:
-            args.extend(self._groupby)
-        if self._apply:
-            args.extend(self._apply)
+        if self._load_statements:
+            args.extend(self._load_statements)
+        if self._groupby_statements:
+            args.extend(self._groupby_statements)
+        if self._apply_statements:
+            args.extend(self._apply_statements)
         if self._sortby_fields:
             sortby_args = []
             for f in self._sortby_fields:

tests/test_search.py

@@ -4095,32 +4095,6 @@ def compare_list_of_dicts(actual, expected):
                     f"All expected:{expected}"
                 )
 
-    @pytest.mark.redismod
-    @skip_if_server_version_lt("8.3.224")
-    def test_review_feedback_hybrid_search(self, client):
-        # Create index and add data
-        self._create_hybrid_search_index(client)
-        self._add_data_for_hybrid_search(client, items_sets=5)
-
-        # set search query
-        search_query = HybridSearchQuery("@color:{red} @color:{green}")
-        search_query.scorer("TFIDF")
-
-        vsim_query = HybridVsimQuery(
-            vector_field_name="@embedding",
-            vector_data=np.array([-100, -200, -200, -300], dtype=np.float32).tobytes(),
-        )
-
-        hybrid_query = HybridQuery(search_query, vsim_query)
-
-        res = client.ft().hybrid_search(query=hybrid_query)
-        if is_resp2_connection(client):
-            assert len(res.results) > 0
-            assert res.warnings == []
-        else:
-            assert len(res["results"]) > 0
-            assert res["warnings"] == []
-
     @pytest.mark.redismod
     @skip_if_server_version_lt("8.3.224")
     def test_basic_hybrid_search(self, client):
@@ -5289,3 +5263,56 @@ def test_hybrid_search_query_with_cursor(self, client):
             assert len(vsim_res_from_cursor.rows) == 5
         else:
             assert len(vsim_res_from_cursor[0]["results"]) == 5
+
+    @pytest.mark.redismod
+    @skip_if_server_version_lt("8.3.224")
+    def test_hybrid_search_query_with_multiple_loads_and_applies(self, client):
+        # Create index and add data
+        self._create_hybrid_search_index(client)
+        self._add_data_for_hybrid_search(client, items_sets=1)
+
+        # set search query
+        search_query = HybridSearchQuery("@color:{red|green}")
+
+        vsim_query = HybridVsimQuery(
+            vector_field_name="@embedding",
+            vector_data=np.array([1, 2, 7, 6], dtype=np.float32).tobytes(),
+        )
+
+        hybrid_query = HybridQuery(search_query, vsim_query)
+
+        postprocessing_config = HybridPostProcessingConfig()
+        postprocessing_config.load("@color", "@price")
+        postprocessing_config.load("@description")
+        postprocessing_config.apply(discount_10_percents="@price - (@price * 0.1)")
+        postprocessing_config.apply(
+            additional_discount="@discount_10_percents - (@discount_10_percents * 0.1)"
+        )
+        postprocessing_config.filter(HybridFilter('@price=="15"'))
+        postprocessing_config.load("@description")
+        postprocessing_config.sort_by(
+            SortbyField("@discount_10_percents", asc=False),
+            SortbyField("@color", asc=True),
+        )
+        postprocessing_config.limit(0, 5)
+
+        res = client.ft().hybrid_search(
+            query=hybrid_query, post_processing=postprocessing_config, timeout=10
+        )
+        print(res)
+        if is_resp2_connection(client):
+            assert len(res.results) == 2
+            for item in res.results:
+                assert item["color"] is not None
+                assert item["price"] is not None
+                assert item["description"] is not None
+                assert item["discount_10_percents"] is not None
+                assert item["additional_discount"] is not None
+        else:
+            assert len(res["results"]) == 2
+            for item in res["results"]:
+                assert item["color"] is not None
+                assert item["price"] is not None
+                assert item["description"] is not None
+                assert item["discount_10_percents"] is not None
+                assert item["additional_discount"] is not None