From f1e1485d6b347b9028e06ba746e7033f9cd5ce3e Mon Sep 17 00:00:00 2001
From: Kishore Nallan <kishorenc@gmail.com>
Date: Tue, 20 Feb 2024 20:45:49 +0530
Subject: [PATCH] Fix nohits query aggregation.

---
 src/collection_manager.cpp       | 20 ++++++------
 test/collection_manager_test.cpp | 55 ++++++++++++++++++++++++++++++++
 test/core_api_utils_test.cpp     |  3 +-
 3 files changed, 66 insertions(+), 12 deletions(-)

diff --git a/src/collection_manager.cpp b/src/collection_manager.cpp
index 5e5a073c..7f1a8aa2 100644
--- a/src/collection_manager.cpp
+++ b/src/collection_manager.cpp
@@ -1837,17 +1837,17 @@ Option<bool> CollectionManager::do_search(std::map<std::string, std::string>& re
     nlohmann::json result = result_op.get();
 
     if(Config::get_instance().get_enable_search_analytics()) {
-        if(result.count("found") != 0 && result["found"].get<size_t>() != 0) {
+        if(result.contains("found")) {
             std::string analytics_query = Tokenizer::normalize_ascii_no_spaces(raw_query);
-            const std::string& expanded_query = Tokenizer::normalize_ascii_no_spaces(
-                                                result["request_params"]["first_q"].get<std::string>());
-
-            AnalyticsManager::get_instance().add_suggestion(orig_coll_name, analytics_query, expanded_query,
-                                                            true, req_params["x-typesense-user-id"]);
-        } else if(result.contains("found") == 0 && result["found"].get<size_t>() == 0) {
-            std::string analytics_query = Tokenizer::normalize_ascii_no_spaces(raw_query);
-            AnalyticsManager::get_instance().add_nohits_query(orig_coll_name, analytics_query,
-                                                              true, req_params["x-typesense-user-id"]);
+            if(result["found"].get<size_t>() != 0) {
+                const std::string& expanded_query = Tokenizer::normalize_ascii_no_spaces(
+                        result["request_params"]["first_q"].get<std::string>());
+                AnalyticsManager::get_instance().add_suggestion(orig_coll_name, analytics_query, expanded_query,
+                                                                true, req_params["x-typesense-user-id"]);
+            } else {
+                AnalyticsManager::get_instance().add_nohits_query(orig_coll_name, analytics_query,
+                                                                  true, req_params["x-typesense-user-id"]);
+            }
         }
     }
 
diff --git a/test/collection_manager_test.cpp b/test/collection_manager_test.cpp
index 538c784d..86482919 100644
--- a/test/collection_manager_test.cpp
+++ b/test/collection_manager_test.cpp
@@ -672,6 +672,61 @@ TEST_F(CollectionManagerTest, QuerySuggestionsShouldBeTrimmed) {
     collectionManager.drop_collection("coll1");
 }
 
+TEST_F(CollectionManagerTest, NoHitsQueryAggregation) {
+    std::vector<field> fields = {field("title", field_types::STRING, false, false, true, "", -1, 1),
+                                 field("year", field_types::INT32, false),
+                                 field("points", field_types::INT32, false),};
+
+    Collection* coll1 = collectionManager.create_collection("coll1", 1, fields, "points").get();
+
+    nlohmann::json doc1;
+    doc1["id"] = "0";
+    doc1["title"] = "Tom Sawyer";
+    doc1["year"] = 1876;
+    doc1["points"] = 100;
+
+    ASSERT_TRUE(coll1->add(doc1.dump()).ok());
+
+    Config::get_instance().set_enable_search_analytics(true);
+
+    nlohmann::json analytics_rule = R"({
+        "name": "nohits_search_queries",
+        "type": "nohits_queries",
+        "params": {
+            "limit": 100,
+            "source": {
+                "collections": ["coll1"]
+            },
+            "destination": {
+                "collection": "nohits_queries"
+            }
+        }
+    })"_json;
+
+    auto create_op = AnalyticsManager::get_instance().create_rule(analytics_rule, false, true);
+    ASSERT_TRUE(create_op.ok());
+
+    nlohmann::json embedded_params;
+    std::map<std::string, std::string> req_params;
+    req_params["collection"] = "coll1";
+    req_params["q"] = "foobarbaz";
+    req_params["query_by"] = "title";
+
+    std::string json_res;
+    auto now_ts = std::chrono::duration_cast<std::chrono::microseconds>(
+            std::chrono::system_clock::now().time_since_epoch()).count();
+
+    auto search_op = collectionManager.do_search(req_params, embedded_params, json_res, now_ts);
+    ASSERT_TRUE(search_op.ok());
+
+    // check that no hits queries have been populated
+    auto nohits_queries = AnalyticsManager::get_instance().get_nohits_queries();
+    ASSERT_EQ(1, nohits_queries["nohits_queries"]->get_user_prefix_queries()[""].size());
+    ASSERT_EQ("foobarbaz", nohits_queries["nohits_queries"]->get_user_prefix_queries()[""][0].query);
+
+    collectionManager.drop_collection("coll1");
+}
+
 TEST_F(CollectionManagerTest, RestoreAutoSchemaDocsOnRestart) {
     Collection *coll1;
 
diff --git a/test/core_api_utils_test.cpp b/test/core_api_utils_test.cpp
index e610ccfc..ba07c356 100644
--- a/test/core_api_utils_test.cpp
+++ b/test/core_api_utils_test.cpp
@@ -3,6 +3,7 @@
 #include <vector>
 #include <collection_manager.h>
 #include <core_api.h>
+#include <analytics_manager.h>
 #include "core_api_utils.h"
 #include "raft_server.h"
 #include "conversation_model_manager.h"
@@ -1051,8 +1052,6 @@ TEST_F(CoreAPIUtilsTest, ExportIncludeExcludeFieldsWithFilter) {
     collectionManager.drop_collection("coll1");
 }
 
-
-
 TEST_F(CoreAPIUtilsTest, TestProxy) {
     std::string res;
     std::unordered_map<std::string, std::string> headers;