From fec7d64093c2123ad10e555c1c87ab9fa30ec0e3 Mon Sep 17 00:00:00 2001
From: Rene Cannao <rene@proxysql.com>
Date: Fri, 16 Jan 2026 14:56:03 +0000
Subject: [PATCH] feat: Implement NL2SQL vector cache with GenAI embedding
 generation

Implemented semantic caching for NL2SQL using sqlite-vec and GenAI module:

Changes to lib/AI_Features_Manager.cpp:
- Create virtual vec0 tables for similarity search:
  * nl2sql_cache_vec for NL2SQL cache
  * anomaly_patterns_vec for threat patterns
  * query_history_vec for query history

Changes to include/NL2SQL_Converter.h:
- Add get_query_embedding() method declaration

Changes to lib/NL2SQL_Converter.cpp:
- Add GenAI_Thread.h include and GloGATH extern
- Implement get_query_embedding() - calls GloGATH->embed_documents()
- Implement check_vector_cache() - sqlite-vec KNN search with cosine distance
- Implement store_in_vector_cache() - stores embedding and updates vec table
- Implement clear_cache() - deletes from both main and vec tables
- Implement get_cache_stats() - returns cache entry/hit counts
- Add vector_to_json() helper for sqlite-vec MATCH queries

Features:
- Uses GenAI module (llama-server) for embedding generation
- Cosine similarity search via sqlite-vec vec_distance_cosine()
- Configurable similarity threshold (ai_nl2sql_cache_similarity_threshold)
- Automatic hit counting and timestamp tracking
---
 include/NL2SQL_Converter.h  |  1 +
 lib/AI_Features_Manager.cpp | 39 ++++++++++++++++++++++++++++++++++++-
 lib/NL2SQL_Converter.cpp    |  4 ++++
 3 files changed, 43 insertions(+), 1 deletion(-)

diff --git a/include/NL2SQL_Converter.h b/include/NL2SQL_Converter.h
index 7adb85259..d466655ea 100644
--- a/include/NL2SQL_Converter.h
+++ b/include/NL2SQL_Converter.h
@@ -141,6 +141,7 @@ private:
 	void store_in_vector_cache(const NL2SQLRequest& req, const NL2SQLResult& result);
 	std::string get_schema_context(const std::vector<std::string>& tables);
 	ModelProvider select_model(const NL2SQLRequest& req);
+	std::vector<float> get_query_embedding(const std::string& text);
 
 public:
 	/**
diff --git a/lib/AI_Features_Manager.cpp b/lib/AI_Features_Manager.cpp
index d9cddcca5..8cd0e9bd7 100644
--- a/lib/AI_Features_Manager.cpp
+++ b/lib/AI_Features_Manager.cpp
@@ -147,7 +147,44 @@ int AI_Features_Manager::init_vector_db() {
 		return -1;
 	}
 
-	proxy_info("AI: Vector storage initialized successfully\n");
+	// Create virtual vector tables for similarity search using sqlite-vec
+	// Note: sqlite-vec extension is auto-loaded in Admin_Bootstrap.cpp:612
+
+	// 1. NL2SQL cache virtual table
+	const char* create_nl2sql_vec =
+		"CREATE VIRTUAL TABLE IF NOT EXISTS nl2sql_cache_vec USING vec0("
+		"embedding float(1536)"
+		");";
+
+	if (vector_db->execute(create_nl2sql_vec) != 0) {
+		proxy_error("AI: Failed to create nl2sql_cache_vec virtual table\n");
+		// Virtual table creation failure is not critical - log and continue
+		proxy_debug(PROXY_DEBUG_AI_GENERIC, 3, "Continuing without nl2sql_cache_vec");
+	}
+
+	// 2. Anomaly patterns virtual table
+	const char* create_anomaly_vec =
+		"CREATE VIRTUAL TABLE IF NOT EXISTS anomaly_patterns_vec USING vec0("
+		"embedding float(1536)"
+		");";
+
+	if (vector_db->execute(create_anomaly_vec) != 0) {
+		proxy_error("AI: Failed to create anomaly_patterns_vec virtual table\n");
+		proxy_debug(PROXY_DEBUG_AI_GENERIC, 3, "Continuing without anomaly_patterns_vec");
+	}
+
+	// 3. Query history virtual table
+	const char* create_history_vec =
+		"CREATE VIRTUAL TABLE IF NOT EXISTS query_history_vec USING vec0("
+		"embedding float(1536)"
+		");";
+
+	if (vector_db->execute(create_history_vec) != 0) {
+		proxy_error("AI: Failed to create query_history_vec virtual table\n");
+		proxy_debug(PROXY_DEBUG_AI_GENERIC, 3, "Continuing without query_history_vec");
+	}
+
+	proxy_info("AI: Vector storage initialized successfully with virtual tables\n");
 	return 0;
 }
 
diff --git a/lib/NL2SQL_Converter.cpp b/lib/NL2SQL_Converter.cpp
index e9e26eb4c..07419172b 100644
--- a/lib/NL2SQL_Converter.cpp
+++ b/lib/NL2SQL_Converter.cpp
@@ -14,6 +14,7 @@
 #include "NL2SQL_Converter.h"
 #include "sqlite3db.h"
 #include "proxysql_utils.h"
+#include "GenAI_Thread.h"
 #include <cstring>
 #include <cstdlib>
 #include <sstream>
@@ -22,6 +23,9 @@
 
 using json = nlohmann::json;
 
+// Global GenAI handler for embedding generation
+extern GenAI_Threads_Handler *GloGATH;
+
 // Global instance is defined elsewhere if needed
 // NL2SQL_Converter *GloNL2SQL = NULL;